Import Packages and Files¶
In [1]:
import torch.nn as nn
from functionalities import dataloader as dl
from functionalities import evaluater as ev
from functionalities import filemanager as fm
from functionalities import trainer as tr
from functionalities import plot as p
from architecture import RotNet as RN
Load Dataset¶
In [ ]:
trainset, testset, classes = dl.load_cifar("./datasets")
In [ ]:
trainloader, validloader, testloader = dl.make_dataloaders(trainset, testset, 128)
Initialize Loss Criterion¶
In [ ]:
criterion = nn.CrossEntropyLoss()
Train RotNet for Rotation Task and Classifiers on Feature Maps¶
In [ ]:
# set rot classes
rot_classes = ['original', '90 rotation', '180 rotation', '270 rotation']
3 Block RotNet¶
In [ ]:
# initialize network
net_block3 = RN.RotNet(num_classes=4, num_conv_block=3, add_avg_pool=False)
In [13]:
# train network
rot_block3_loss_log, _, rot_block3_test_accuracy_log, _, _ = tr.adaptive_learning([0.1, 0.02, 0.004, 0.0008],
[60, 120, 160, 200], 0.9, 5e-4, net_block3, criterion, trainloader, None, testloader, rot=['90', '180', '270'])
[1, 60] loss: 1.141 [1, 120] loss: 0.999 [1, 180] loss: 0.920 [1, 240] loss: 0.851 [1, 300] loss: 0.790 [1, 360] loss: 0.759 Epoch: 1 -> Loss: 0.787678480148 Epoch: 1 -> Test Accuracy: 69.21 [2, 60] loss: 0.701 [2, 120] loss: 0.696 [2, 180] loss: 0.677 [2, 240] loss: 0.648 [2, 300] loss: 0.647 [2, 360] loss: 0.616 Epoch: 2 -> Loss: 0.662011623383 Epoch: 2 -> Test Accuracy: 76.055 [3, 60] loss: 0.596 [3, 120] loss: 0.581 [3, 180] loss: 0.577 [3, 240] loss: 0.583 [3, 300] loss: 0.557 [3, 360] loss: 0.566 Epoch: 3 -> Loss: 0.581205248833 Epoch: 3 -> Test Accuracy: 78.8075 [4, 60] loss: 0.544 [4, 120] loss: 0.519 [4, 180] loss: 0.517 [4, 240] loss: 0.537 [4, 300] loss: 0.507 [4, 360] loss: 0.507 Epoch: 4 -> Loss: 0.558216452599 Epoch: 4 -> Test Accuracy: 79.01 [5, 60] loss: 0.496 [5, 120] loss: 0.497 [5, 180] loss: 0.486 [5, 240] loss: 0.495 [5, 300] loss: 0.473 [5, 360] loss: 0.472 Epoch: 5 -> Loss: 0.537355840206 Epoch: 5 -> Test Accuracy: 81.725 [6, 60] loss: 0.449 [6, 120] loss: 0.464 [6, 180] loss: 0.462 [6, 240] loss: 0.465 [6, 300] loss: 0.466 [6, 360] loss: 0.452 Epoch: 6 -> Loss: 0.470865309238 Epoch: 6 -> Test Accuracy: 81.4725 [7, 60] loss: 0.439 [7, 120] loss: 0.433 [7, 180] loss: 0.429 [7, 240] loss: 0.451 [7, 300] loss: 0.439 [7, 360] loss: 0.450 Epoch: 7 -> Loss: 0.556775391102 Epoch: 7 -> Test Accuracy: 83.0 [8, 60] loss: 0.432 [8, 120] loss: 0.434 [8, 180] loss: 0.416 [8, 240] loss: 0.422 [8, 300] loss: 0.417 [8, 360] loss: 0.429 Epoch: 8 -> Loss: 0.499394506216 Epoch: 8 -> Test Accuracy: 82.2925 [9, 60] loss: 0.406 [9, 120] loss: 0.397 [9, 180] loss: 0.410 [9, 240] loss: 0.417 [9, 300] loss: 0.413 [9, 360] loss: 0.406 Epoch: 9 -> Loss: 0.406752169132 Epoch: 9 -> Test Accuracy: 82.8125 [10, 60] loss: 0.386 [10, 120] loss: 0.398 [10, 180] loss: 0.404 [10, 240] loss: 0.402 [10, 300] loss: 0.401 [10, 360] loss: 0.402 Epoch: 10 -> Loss: 0.341294229031 Epoch: 10 -> Test Accuracy: 84.595 [11, 60] loss: 0.385 [11, 120] loss: 0.395 [11, 180] loss: 0.400 [11, 240] loss: 0.379 [11, 300] loss: 0.391 [11, 360] loss: 0.401 Epoch: 11 -> Loss: 0.273233801126 Epoch: 11 -> Test Accuracy: 84.7425 [12, 60] loss: 0.388 [12, 120] loss: 0.380 [12, 180] loss: 0.389 [12, 240] loss: 0.374 [12, 300] loss: 0.375 [12, 360] loss: 0.389 Epoch: 12 -> Loss: 0.283909648657 Epoch: 12 -> Test Accuracy: 84.285 [13, 60] loss: 0.380 [13, 120] loss: 0.366 [13, 180] loss: 0.369 [13, 240] loss: 0.375 [13, 300] loss: 0.378 [13, 360] loss: 0.368 Epoch: 13 -> Loss: 0.423889309168 Epoch: 13 -> Test Accuracy: 84.8575 [14, 60] loss: 0.345 [14, 120] loss: 0.374 [14, 180] loss: 0.375 [14, 240] loss: 0.373 [14, 300] loss: 0.385 [14, 360] loss: 0.377 Epoch: 14 -> Loss: 0.427053511143 Epoch: 14 -> Test Accuracy: 85.68 [15, 60] loss: 0.357 [15, 120] loss: 0.350 [15, 180] loss: 0.352 [15, 240] loss: 0.354 [15, 300] loss: 0.373 [15, 360] loss: 0.377 Epoch: 15 -> Loss: 0.273929357529 Epoch: 15 -> Test Accuracy: 85.0625 [16, 60] loss: 0.342 [16, 120] loss: 0.355 [16, 180] loss: 0.359 [16, 240] loss: 0.362 [16, 300] loss: 0.356 [16, 360] loss: 0.351 Epoch: 16 -> Loss: 0.370511502028 Epoch: 16 -> Test Accuracy: 85.84 [17, 60] loss: 0.340 [17, 120] loss: 0.351 [17, 180] loss: 0.350 [17, 240] loss: 0.339 [17, 300] loss: 0.372 [17, 360] loss: 0.364 Epoch: 17 -> Loss: 0.440175831318 Epoch: 17 -> Test Accuracy: 85.1375 [18, 60] loss: 0.346 [18, 120] loss: 0.324 [18, 180] loss: 0.364 [18, 240] loss: 0.361 [18, 300] loss: 0.339 [18, 360] loss: 0.349 Epoch: 18 -> Loss: 0.436338275671 Epoch: 18 -> Test Accuracy: 84.84 [19, 60] loss: 0.345 [19, 120] loss: 0.332 [19, 180] loss: 0.351 [19, 240] loss: 0.348 [19, 300] loss: 0.347 [19, 360] loss: 0.353 Epoch: 19 -> Loss: 0.369504094124 Epoch: 19 -> Test Accuracy: 85.5425 [20, 60] loss: 0.332 [20, 120] loss: 0.338 [20, 180] loss: 0.358 [20, 240] loss: 0.338 [20, 300] loss: 0.341 [20, 360] loss: 0.350 Epoch: 20 -> Loss: 0.225799173117 Epoch: 20 -> Test Accuracy: 86.325 [21, 60] loss: 0.337 [21, 120] loss: 0.341 [21, 180] loss: 0.337 [21, 240] loss: 0.341 [21, 300] loss: 0.334 [21, 360] loss: 0.337 Epoch: 21 -> Loss: 0.392786115408 Epoch: 21 -> Test Accuracy: 86.0875 [22, 60] loss: 0.336 [22, 120] loss: 0.334 [22, 180] loss: 0.334 [22, 240] loss: 0.339 [22, 300] loss: 0.337 [22, 360] loss: 0.330 Epoch: 22 -> Loss: 0.356912195683 Epoch: 22 -> Test Accuracy: 85.8025 [23, 60] loss: 0.309 [23, 120] loss: 0.337 [23, 180] loss: 0.343 [23, 240] loss: 0.327 [23, 300] loss: 0.342 [23, 360] loss: 0.333 Epoch: 23 -> Loss: 0.468298614025 Epoch: 23 -> Test Accuracy: 85.7675 [24, 60] loss: 0.332 [24, 120] loss: 0.321 [24, 180] loss: 0.338 [24, 240] loss: 0.338 [24, 300] loss: 0.337 [24, 360] loss: 0.332 Epoch: 24 -> Loss: 0.208731681108 Epoch: 24 -> Test Accuracy: 86.345 [25, 60] loss: 0.330 [25, 120] loss: 0.324 [25, 180] loss: 0.336 [25, 240] loss: 0.326 [25, 300] loss: 0.336 [25, 360] loss: 0.329 Epoch: 25 -> Loss: 0.385006994009 Epoch: 25 -> Test Accuracy: 85.9125 [26, 60] loss: 0.318 [26, 120] loss: 0.318 [26, 180] loss: 0.328 [26, 240] loss: 0.324 [26, 300] loss: 0.331 [26, 360] loss: 0.334 Epoch: 26 -> Loss: 0.294116735458 Epoch: 26 -> Test Accuracy: 86.9775 [27, 60] loss: 0.305 [27, 120] loss: 0.330 [27, 180] loss: 0.327 [27, 240] loss: 0.331 [27, 300] loss: 0.339 [27, 360] loss: 0.319 Epoch: 27 -> Loss: 0.295988678932 Epoch: 27 -> Test Accuracy: 86.8975 [28, 60] loss: 0.330 [28, 120] loss: 0.316 [28, 180] loss: 0.325 [28, 240] loss: 0.324 [28, 300] loss: 0.318 [28, 360] loss: 0.342 Epoch: 28 -> Loss: 0.233337074518 Epoch: 28 -> Test Accuracy: 86.135 [29, 60] loss: 0.317 [29, 120] loss: 0.330 [29, 180] loss: 0.321 [29, 240] loss: 0.322 [29, 300] loss: 0.332 [29, 360] loss: 0.313 Epoch: 29 -> Loss: 0.352734535933 Epoch: 29 -> Test Accuracy: 86.1825 [30, 60] loss: 0.299 [30, 120] loss: 0.326 [30, 180] loss: 0.321 [30, 240] loss: 0.328 [30, 300] loss: 0.327 [30, 360] loss: 0.329 Epoch: 30 -> Loss: 0.335489243269 Epoch: 30 -> Test Accuracy: 86.6575 [31, 60] loss: 0.306 [31, 120] loss: 0.327 [31, 180] loss: 0.323 [31, 240] loss: 0.322 [31, 300] loss: 0.326 [31, 360] loss: 0.315 Epoch: 31 -> Loss: 0.512306988239 Epoch: 31 -> Test Accuracy: 85.37 [32, 60] loss: 0.297 [32, 120] loss: 0.316 [32, 180] loss: 0.340 [32, 240] loss: 0.321 [32, 300] loss: 0.310 [32, 360] loss: 0.321 Epoch: 32 -> Loss: 0.35365241766 Epoch: 32 -> Test Accuracy: 86.9825 [33, 60] loss: 0.319 [33, 120] loss: 0.318 [33, 180] loss: 0.328 [33, 240] loss: 0.319 [33, 300] loss: 0.313 [33, 360] loss: 0.321 Epoch: 33 -> Loss: 0.385478198528 Epoch: 33 -> Test Accuracy: 86.2025 [34, 60] loss: 0.318 [34, 120] loss: 0.316 [34, 180] loss: 0.300 [34, 240] loss: 0.324 [34, 300] loss: 0.319 [34, 360] loss: 0.311 Epoch: 34 -> Loss: 0.343362241983 Epoch: 34 -> Test Accuracy: 85.1975 [35, 60] loss: 0.309 [35, 120] loss: 0.325 [35, 180] loss: 0.327 [35, 240] loss: 0.322 [35, 300] loss: 0.320 [35, 360] loss: 0.320 Epoch: 35 -> Loss: 0.272876352072 Epoch: 35 -> Test Accuracy: 86.99 [36, 60] loss: 0.307 [36, 120] loss: 0.308 [36, 180] loss: 0.311 [36, 240] loss: 0.322 [36, 300] loss: 0.314 [36, 360] loss: 0.322 Epoch: 36 -> Loss: 0.267319113016 Epoch: 36 -> Test Accuracy: 86.2175 [37, 60] loss: 0.304 [37, 120] loss: 0.319 [37, 180] loss: 0.311 [37, 240] loss: 0.317 [37, 300] loss: 0.317 [37, 360] loss: 0.314 Epoch: 37 -> Loss: 0.215385004878 Epoch: 37 -> Test Accuracy: 87.205 [38, 60] loss: 0.300 [38, 120] loss: 0.300 [38, 180] loss: 0.300 [38, 240] loss: 0.317 [38, 300] loss: 0.318 [38, 360] loss: 0.315 Epoch: 38 -> Loss: 0.333074420691 Epoch: 38 -> Test Accuracy: 86.69 [39, 60] loss: 0.298 [39, 120] loss: 0.313 [39, 180] loss: 0.312 [39, 240] loss: 0.314 [39, 300] loss: 0.312 [39, 360] loss: 0.322 Epoch: 39 -> Loss: 0.329833418131 Epoch: 39 -> Test Accuracy: 87.2125 [40, 60] loss: 0.298 [40, 120] loss: 0.299 [40, 180] loss: 0.309 [40, 240] loss: 0.317 [40, 300] loss: 0.317 [40, 360] loss: 0.316 Epoch: 40 -> Loss: 0.324939012527 Epoch: 40 -> Test Accuracy: 87.0475 [41, 60] loss: 0.303 [41, 120] loss: 0.304 [41, 180] loss: 0.305 [41, 240] loss: 0.304 [41, 300] loss: 0.327 [41, 360] loss: 0.309 Epoch: 41 -> Loss: 0.334853470325 Epoch: 41 -> Test Accuracy: 86.5025 [42, 60] loss: 0.309 [42, 120] loss: 0.306 [42, 180] loss: 0.314 [42, 240] loss: 0.322 [42, 300] loss: 0.311 [42, 360] loss: 0.313 Epoch: 42 -> Loss: 0.318148314953 Epoch: 42 -> Test Accuracy: 87.27 [43, 60] loss: 0.284 [43, 120] loss: 0.312 [43, 180] loss: 0.322 [43, 240] loss: 0.291 [43, 300] loss: 0.321 [43, 360] loss: 0.304 Epoch: 43 -> Loss: 0.382598012686 Epoch: 43 -> Test Accuracy: 83.9325 [44, 60] loss: 0.311 [44, 120] loss: 0.305 [44, 180] loss: 0.303 [44, 240] loss: 0.298 [44, 300] loss: 0.310 [44, 360] loss: 0.330 Epoch: 44 -> Loss: 0.220706671476 Epoch: 44 -> Test Accuracy: 87.075 [45, 60] loss: 0.298 [45, 120] loss: 0.296 [45, 180] loss: 0.313 [45, 240] loss: 0.317 [45, 300] loss: 0.302 [45, 360] loss: 0.308 Epoch: 45 -> Loss: 0.32049909234 Epoch: 45 -> Test Accuracy: 87.4575 [46, 60] loss: 0.290 [46, 120] loss: 0.306 [46, 180] loss: 0.302 [46, 240] loss: 0.307 [46, 300] loss: 0.309 [46, 360] loss: 0.316 Epoch: 46 -> Loss: 0.323618233204 Epoch: 46 -> Test Accuracy: 87.1525 [47, 60] loss: 0.294 [47, 120] loss: 0.285 [47, 180] loss: 0.300 [47, 240] loss: 0.322 [47, 300] loss: 0.311 [47, 360] loss: 0.307 Epoch: 47 -> Loss: 0.383451044559 Epoch: 47 -> Test Accuracy: 86.2075 [48, 60] loss: 0.297 [48, 120] loss: 0.300 [48, 180] loss: 0.293 [48, 240] loss: 0.324 [48, 300] loss: 0.312 [48, 360] loss: 0.307 Epoch: 48 -> Loss: 0.247455790639 Epoch: 48 -> Test Accuracy: 87.21 [49, 60] loss: 0.280 [49, 120] loss: 0.305 [49, 180] loss: 0.303 [49, 240] loss: 0.310 [49, 300] loss: 0.311 [49, 360] loss: 0.316 Epoch: 49 -> Loss: 0.177563875914 Epoch: 49 -> Test Accuracy: 86.7775 [50, 60] loss: 0.310 [50, 120] loss: 0.299 [50, 180] loss: 0.295 [50, 240] loss: 0.303 [50, 300] loss: 0.308 [50, 360] loss: 0.317 Epoch: 50 -> Loss: 0.344310581684 Epoch: 50 -> Test Accuracy: 87.055 [51, 60] loss: 0.287 [51, 120] loss: 0.293 [51, 180] loss: 0.309 [51, 240] loss: 0.303 [51, 300] loss: 0.308 [51, 360] loss: 0.311 Epoch: 51 -> Loss: 0.226258903742 Epoch: 51 -> Test Accuracy: 87.18 [52, 60] loss: 0.288 [52, 120] loss: 0.309 [52, 180] loss: 0.302 [52, 240] loss: 0.311 [52, 300] loss: 0.306 [52, 360] loss: 0.306 Epoch: 52 -> Loss: 0.207420393825 Epoch: 52 -> Test Accuracy: 85.79 [53, 60] loss: 0.296 [53, 120] loss: 0.301 [53, 180] loss: 0.297 [53, 240] loss: 0.297 [53, 300] loss: 0.304 [53, 360] loss: 0.310 Epoch: 53 -> Loss: 0.364464432001 Epoch: 53 -> Test Accuracy: 87.105 [54, 60] loss: 0.300 [54, 120] loss: 0.304 [54, 180] loss: 0.313 [54, 240] loss: 0.303 [54, 300] loss: 0.302 [54, 360] loss: 0.298 Epoch: 54 -> Loss: 0.423810869455 Epoch: 54 -> Test Accuracy: 86.735 [55, 60] loss: 0.292 [55, 120] loss: 0.302 [55, 180] loss: 0.307 [55, 240] loss: 0.300 [55, 300] loss: 0.294 [55, 360] loss: 0.310 Epoch: 55 -> Loss: 0.418384850025 Epoch: 55 -> Test Accuracy: 87.13 [56, 60] loss: 0.289 [56, 120] loss: 0.310 [56, 180] loss: 0.288 [56, 240] loss: 0.301 [56, 300] loss: 0.298 [56, 360] loss: 0.308 Epoch: 56 -> Loss: 0.383456289768 Epoch: 56 -> Test Accuracy: 86.5475 [57, 60] loss: 0.295 [57, 120] loss: 0.302 [57, 180] loss: 0.305 [57, 240] loss: 0.306 [57, 300] loss: 0.297 [57, 360] loss: 0.298 Epoch: 57 -> Loss: 0.171087294817 Epoch: 57 -> Test Accuracy: 87.1125 [58, 60] loss: 0.280 [58, 120] loss: 0.290 [58, 180] loss: 0.306 [58, 240] loss: 0.300 [58, 300] loss: 0.307 [58, 360] loss: 0.291 Epoch: 58 -> Loss: 0.33591324091 Epoch: 58 -> Test Accuracy: 86.4525 [59, 60] loss: 0.285 [59, 120] loss: 0.292 [59, 180] loss: 0.299 [59, 240] loss: 0.301 [59, 300] loss: 0.307 [59, 360] loss: 0.308 Epoch: 59 -> Loss: 0.359701931477 Epoch: 59 -> Test Accuracy: 86.6825 [60, 60] loss: 0.296 [60, 120] loss: 0.296 [60, 180] loss: 0.294 [60, 240] loss: 0.295 [60, 300] loss: 0.304 [60, 360] loss: 0.283 Epoch: 60 -> Loss: 0.189138680696 Epoch: 60 -> Test Accuracy: 86.9475 [61, 60] loss: 0.233 [61, 120] loss: 0.200 [61, 180] loss: 0.192 [61, 240] loss: 0.179 [61, 300] loss: 0.185 [61, 360] loss: 0.176 Epoch: 61 -> Loss: 0.169828921556 Epoch: 61 -> Test Accuracy: 91.1325 [62, 60] loss: 0.152 [62, 120] loss: 0.171 [62, 180] loss: 0.173 [62, 240] loss: 0.170 [62, 300] loss: 0.175 [62, 360] loss: 0.172 Epoch: 62 -> Loss: 0.253974795341 Epoch: 62 -> Test Accuracy: 91.3325 [63, 60] loss: 0.144 [63, 120] loss: 0.151 [63, 180] loss: 0.154 [63, 240] loss: 0.161 [63, 300] loss: 0.159 [63, 360] loss: 0.156 Epoch: 63 -> Loss: 0.121799066663 Epoch: 63 -> Test Accuracy: 91.0675 [64, 60] loss: 0.147 [64, 120] loss: 0.150 [64, 180] loss: 0.157 [64, 240] loss: 0.149 [64, 300] loss: 0.159 [64, 360] loss: 0.152 Epoch: 64 -> Loss: 0.203890949488 Epoch: 64 -> Test Accuracy: 91.1125 [65, 60] loss: 0.148 [65, 120] loss: 0.141 [65, 180] loss: 0.142 [65, 240] loss: 0.159 [65, 300] loss: 0.150 [65, 360] loss: 0.143 Epoch: 65 -> Loss: 0.0973534584045 Epoch: 65 -> Test Accuracy: 91.3175 [66, 60] loss: 0.138 [66, 120] loss: 0.137 [66, 180] loss: 0.146 [66, 240] loss: 0.144 [66, 300] loss: 0.156 [66, 360] loss: 0.154 Epoch: 66 -> Loss: 0.157145500183 Epoch: 66 -> Test Accuracy: 90.9675 [67, 60] loss: 0.140 [67, 120] loss: 0.142 [67, 180] loss: 0.138 [67, 240] loss: 0.143 [67, 300] loss: 0.155 [67, 360] loss: 0.150 Epoch: 67 -> Loss: 0.144269049168 Epoch: 67 -> Test Accuracy: 91.22 [68, 60] loss: 0.138 [68, 120] loss: 0.141 [68, 180] loss: 0.138 [68, 240] loss: 0.142 [68, 300] loss: 0.149 [68, 360] loss: 0.156 Epoch: 68 -> Loss: 0.115117274225 Epoch: 68 -> Test Accuracy: 90.9275 [69, 60] loss: 0.130 [69, 120] loss: 0.150 [69, 180] loss: 0.140 [69, 240] loss: 0.147 [69, 300] loss: 0.149 [69, 360] loss: 0.151 Epoch: 69 -> Loss: 0.166590347886 Epoch: 69 -> Test Accuracy: 91.17 [70, 60] loss: 0.134 [70, 120] loss: 0.135 [70, 180] loss: 0.146 [70, 240] loss: 0.142 [70, 300] loss: 0.153 [70, 360] loss: 0.151 Epoch: 70 -> Loss: 0.12411685288 Epoch: 70 -> Test Accuracy: 91.2375 [71, 60] loss: 0.135 [71, 120] loss: 0.138 [71, 180] loss: 0.142 [71, 240] loss: 0.148 [71, 300] loss: 0.149 [71, 360] loss: 0.150 Epoch: 71 -> Loss: 0.111194655299 Epoch: 71 -> Test Accuracy: 90.845 [72, 60] loss: 0.138 [72, 120] loss: 0.142 [72, 180] loss: 0.139 [72, 240] loss: 0.145 [72, 300] loss: 0.151 [72, 360] loss: 0.158 Epoch: 72 -> Loss: 0.115612111986 Epoch: 72 -> Test Accuracy: 90.725 [73, 60] loss: 0.136 [73, 120] loss: 0.148 [73, 180] loss: 0.146 [73, 240] loss: 0.137 [73, 300] loss: 0.145 [73, 360] loss: 0.153 Epoch: 73 -> Loss: 0.173922792077 Epoch: 73 -> Test Accuracy: 90.495 [74, 60] loss: 0.137 [74, 120] loss: 0.132 [74, 180] loss: 0.148 [74, 240] loss: 0.156 [74, 300] loss: 0.160 [74, 360] loss: 0.153 Epoch: 74 -> Loss: 0.161368072033 Epoch: 74 -> Test Accuracy: 90.6275 [75, 60] loss: 0.137 [75, 120] loss: 0.132 [75, 180] loss: 0.149 [75, 240] loss: 0.152 [75, 300] loss: 0.149 [75, 360] loss: 0.144 Epoch: 75 -> Loss: 0.264967113733 Epoch: 75 -> Test Accuracy: 90.45 [76, 60] loss: 0.135 [76, 120] loss: 0.142 [76, 180] loss: 0.146 [76, 240] loss: 0.158 [76, 300] loss: 0.145 [76, 360] loss: 0.151 Epoch: 76 -> Loss: 0.18966922164 Epoch: 76 -> Test Accuracy: 90.3475 [77, 60] loss: 0.134 [77, 120] loss: 0.135 [77, 180] loss: 0.137 [77, 240] loss: 0.144 [77, 300] loss: 0.159 [77, 360] loss: 0.164 Epoch: 77 -> Loss: 0.165032312274 Epoch: 77 -> Test Accuracy: 89.9725 [78, 60] loss: 0.137 [78, 120] loss: 0.137 [78, 180] loss: 0.150 [78, 240] loss: 0.148 [78, 300] loss: 0.148 [78, 360] loss: 0.147 Epoch: 78 -> Loss: 0.118702635169 Epoch: 78 -> Test Accuracy: 90.21 [79, 60] loss: 0.142 [79, 120] loss: 0.139 [79, 180] loss: 0.141 [79, 240] loss: 0.145 [79, 300] loss: 0.145 [79, 360] loss: 0.157 Epoch: 79 -> Loss: 0.200592786074 Epoch: 79 -> Test Accuracy: 90.805 [80, 60] loss: 0.146 [80, 120] loss: 0.140 [80, 180] loss: 0.146 [80, 240] loss: 0.144 [80, 300] loss: 0.146 [80, 360] loss: 0.163 Epoch: 80 -> Loss: 0.0748644471169 Epoch: 80 -> Test Accuracy: 90.3125 [81, 60] loss: 0.147 [81, 120] loss: 0.146 [81, 180] loss: 0.151 [81, 240] loss: 0.139 [81, 300] loss: 0.150 [81, 360] loss: 0.157 Epoch: 81 -> Loss: 0.112988092005 Epoch: 81 -> Test Accuracy: 90.3075 [82, 60] loss: 0.138 [82, 120] loss: 0.138 [82, 180] loss: 0.152 [82, 240] loss: 0.144 [82, 300] loss: 0.144 [82, 360] loss: 0.154 Epoch: 82 -> Loss: 0.0915526524186 Epoch: 82 -> Test Accuracy: 90.405 [83, 60] loss: 0.130 [83, 120] loss: 0.140 [83, 180] loss: 0.146 [83, 240] loss: 0.145 [83, 300] loss: 0.153 [83, 360] loss: 0.158 Epoch: 83 -> Loss: 0.156742066145 Epoch: 83 -> Test Accuracy: 90.6825 [84, 60] loss: 0.135 [84, 120] loss: 0.139 [84, 180] loss: 0.137 [84, 240] loss: 0.145 [84, 300] loss: 0.158 [84, 360] loss: 0.149 Epoch: 84 -> Loss: 0.114132240415 Epoch: 84 -> Test Accuracy: 90.4775 [85, 60] loss: 0.129 [85, 120] loss: 0.129 [85, 180] loss: 0.137 [85, 240] loss: 0.145 [85, 300] loss: 0.148 [85, 360] loss: 0.152 Epoch: 85 -> Loss: 0.170121192932 Epoch: 85 -> Test Accuracy: 90.3325 [86, 60] loss: 0.132 [86, 120] loss: 0.145 [86, 180] loss: 0.141 [86, 240] loss: 0.146 [86, 300] loss: 0.144 [86, 360] loss: 0.155 Epoch: 86 -> Loss: 0.139373719692 Epoch: 86 -> Test Accuracy: 90.41 [87, 60] loss: 0.136 [87, 120] loss: 0.136 [87, 180] loss: 0.147 [87, 240] loss: 0.148 [87, 300] loss: 0.145 [87, 360] loss: 0.157 Epoch: 87 -> Loss: 0.234496861696 Epoch: 87 -> Test Accuracy: 89.7625 [88, 60] loss: 0.134 [88, 120] loss: 0.139 [88, 180] loss: 0.140 [88, 240] loss: 0.139 [88, 300] loss: 0.155 [88, 360] loss: 0.162 Epoch: 88 -> Loss: 0.169020205736 Epoch: 88 -> Test Accuracy: 90.3625 [89, 60] loss: 0.132 [89, 120] loss: 0.142 [89, 180] loss: 0.144 [89, 240] loss: 0.148 [89, 300] loss: 0.141 [89, 360] loss: 0.141 Epoch: 89 -> Loss: 0.257692873478 Epoch: 89 -> Test Accuracy: 90.35 [90, 60] loss: 0.135 [90, 120] loss: 0.136 [90, 180] loss: 0.148 [90, 240] loss: 0.139 [90, 300] loss: 0.152 [90, 360] loss: 0.150 Epoch: 90 -> Loss: 0.183051347733 Epoch: 90 -> Test Accuracy: 90.4525 [91, 60] loss: 0.127 [91, 120] loss: 0.136 [91, 180] loss: 0.146 [91, 240] loss: 0.146 [91, 300] loss: 0.143 [91, 360] loss: 0.146 Epoch: 91 -> Loss: 0.230325505137 Epoch: 91 -> Test Accuracy: 90.575 [92, 60] loss: 0.137 [92, 120] loss: 0.130 [92, 180] loss: 0.133 [92, 240] loss: 0.142 [92, 300] loss: 0.144 [92, 360] loss: 0.150 Epoch: 92 -> Loss: 0.177278190851 Epoch: 92 -> Test Accuracy: 90.1125 [93, 60] loss: 0.124 [93, 120] loss: 0.132 [93, 180] loss: 0.143 [93, 240] loss: 0.145 [93, 300] loss: 0.138 [93, 360] loss: 0.150 Epoch: 93 -> Loss: 0.115009739995 Epoch: 93 -> Test Accuracy: 90.295 [94, 60] loss: 0.132 [94, 120] loss: 0.133 [94, 180] loss: 0.134 [94, 240] loss: 0.152 [94, 300] loss: 0.152 [94, 360] loss: 0.144 Epoch: 94 -> Loss: 0.123808979988 Epoch: 94 -> Test Accuracy: 90.675 [95, 60] loss: 0.140 [95, 120] loss: 0.130 [95, 180] loss: 0.147 [95, 240] loss: 0.143 [95, 300] loss: 0.146 [95, 360] loss: 0.145 Epoch: 95 -> Loss: 0.0961346998811 Epoch: 95 -> Test Accuracy: 90.2625 [96, 60] loss: 0.130 [96, 120] loss: 0.129 [96, 180] loss: 0.139 [96, 240] loss: 0.147 [96, 300] loss: 0.152 [96, 360] loss: 0.144 Epoch: 96 -> Loss: 0.0893204286695 Epoch: 96 -> Test Accuracy: 90.52 [97, 60] loss: 0.124 [97, 120] loss: 0.135 [97, 180] loss: 0.141 [97, 240] loss: 0.137 [97, 300] loss: 0.133 [97, 360] loss: 0.146 Epoch: 97 -> Loss: 0.141617566347 Epoch: 97 -> Test Accuracy: 91.06 [98, 60] loss: 0.125 [98, 120] loss: 0.140 [98, 180] loss: 0.143 [98, 240] loss: 0.136 [98, 300] loss: 0.143 [98, 360] loss: 0.150 Epoch: 98 -> Loss: 0.166911140084 Epoch: 98 -> Test Accuracy: 90.305 [99, 60] loss: 0.129 [99, 120] loss: 0.130 [99, 180] loss: 0.136 [99, 240] loss: 0.139 [99, 300] loss: 0.149 [99, 360] loss: 0.140 Epoch: 99 -> Loss: 0.151396125555 Epoch: 99 -> Test Accuracy: 89.95 [100, 60] loss: 0.130 [100, 120] loss: 0.133 [100, 180] loss: 0.135 [100, 240] loss: 0.144 [100, 300] loss: 0.147 [100, 360] loss: 0.154 Epoch: 100 -> Loss: 0.127764731646 Epoch: 100 -> Test Accuracy: 90.2675 [101, 60] loss: 0.135 [101, 120] loss: 0.125 [101, 180] loss: 0.139 [101, 240] loss: 0.136 [101, 300] loss: 0.151 [101, 360] loss: 0.141 Epoch: 101 -> Loss: 0.155829519033 Epoch: 101 -> Test Accuracy: 90.58 [102, 60] loss: 0.127 [102, 120] loss: 0.132 [102, 180] loss: 0.140 [102, 240] loss: 0.142 [102, 300] loss: 0.142 [102, 360] loss: 0.148 Epoch: 102 -> Loss: 0.149779215455 Epoch: 102 -> Test Accuracy: 90.5275 [103, 60] loss: 0.123 [103, 120] loss: 0.132 [103, 180] loss: 0.137 [103, 240] loss: 0.136 [103, 300] loss: 0.139 [103, 360] loss: 0.144 Epoch: 103 -> Loss: 0.143297225237 Epoch: 103 -> Test Accuracy: 89.9525 [104, 60] loss: 0.137 [104, 120] loss: 0.121 [104, 180] loss: 0.140 [104, 240] loss: 0.140 [104, 300] loss: 0.133 [104, 360] loss: 0.147 Epoch: 104 -> Loss: 0.165147423744 Epoch: 104 -> Test Accuracy: 90.55 [105, 60] loss: 0.127 [105, 120] loss: 0.120 [105, 180] loss: 0.136 [105, 240] loss: 0.147 [105, 300] loss: 0.143 [105, 360] loss: 0.145 Epoch: 105 -> Loss: 0.153953403234 Epoch: 105 -> Test Accuracy: 90.44 [106, 60] loss: 0.120 [106, 120] loss: 0.128 [106, 180] loss: 0.131 [106, 240] loss: 0.136 [106, 300] loss: 0.137 [106, 360] loss: 0.139 Epoch: 106 -> Loss: 0.160528451204 Epoch: 106 -> Test Accuracy: 91.0075 [107, 60] loss: 0.127 [107, 120] loss: 0.121 [107, 180] loss: 0.132 [107, 240] loss: 0.139 [107, 300] loss: 0.142 [107, 360] loss: 0.142 Epoch: 107 -> Loss: 0.0822361558676 Epoch: 107 -> Test Accuracy: 90.775 [108, 60] loss: 0.134 [108, 120] loss: 0.134 [108, 180] loss: 0.135 [108, 240] loss: 0.143 [108, 300] loss: 0.129 [108, 360] loss: 0.143 Epoch: 108 -> Loss: 0.14577370882 Epoch: 108 -> Test Accuracy: 90.03 [109, 60] loss: 0.122 [109, 120] loss: 0.129 [109, 180] loss: 0.130 [109, 240] loss: 0.133 [109, 300] loss: 0.133 [109, 360] loss: 0.149 Epoch: 109 -> Loss: 0.171248614788 Epoch: 109 -> Test Accuracy: 90.3875 [110, 60] loss: 0.128 [110, 120] loss: 0.130 [110, 180] loss: 0.142 [110, 240] loss: 0.139 [110, 300] loss: 0.136 [110, 360] loss: 0.143 Epoch: 110 -> Loss: 0.190067365766 Epoch: 110 -> Test Accuracy: 90.57 [111, 60] loss: 0.121 [111, 120] loss: 0.133 [111, 180] loss: 0.132 [111, 240] loss: 0.138 [111, 300] loss: 0.141 [111, 360] loss: 0.134 Epoch: 111 -> Loss: 0.136060848832 Epoch: 111 -> Test Accuracy: 90.3075 [112, 60] loss: 0.129 [112, 120] loss: 0.127 [112, 180] loss: 0.136 [112, 240] loss: 0.134 [112, 300] loss: 0.141 [112, 360] loss: 0.135 Epoch: 112 -> Loss: 0.13542817533 Epoch: 112 -> Test Accuracy: 90.8225 [113, 60] loss: 0.117 [113, 120] loss: 0.127 [113, 180] loss: 0.133 [113, 240] loss: 0.142 [113, 300] loss: 0.142 [113, 360] loss: 0.131 Epoch: 113 -> Loss: 0.221937775612 Epoch: 113 -> Test Accuracy: 89.9675 [114, 60] loss: 0.129 [114, 120] loss: 0.131 [114, 180] loss: 0.135 [114, 240] loss: 0.133 [114, 300] loss: 0.139 [114, 360] loss: 0.143 Epoch: 114 -> Loss: 0.0751198902726 Epoch: 114 -> Test Accuracy: 90.23 [115, 60] loss: 0.126 [115, 120] loss: 0.116 [115, 180] loss: 0.129 [115, 240] loss: 0.137 [115, 300] loss: 0.134 [115, 360] loss: 0.141 Epoch: 115 -> Loss: 0.137918055058 Epoch: 115 -> Test Accuracy: 90.8375 [116, 60] loss: 0.124 [116, 120] loss: 0.134 [116, 180] loss: 0.136 [116, 240] loss: 0.128 [116, 300] loss: 0.131 [116, 360] loss: 0.139 Epoch: 116 -> Loss: 0.073376826942 Epoch: 116 -> Test Accuracy: 90.7075 [117, 60] loss: 0.121 [117, 120] loss: 0.128 [117, 180] loss: 0.124 [117, 240] loss: 0.139 [117, 300] loss: 0.137 [117, 360] loss: 0.143 Epoch: 117 -> Loss: 0.135540992022 Epoch: 117 -> Test Accuracy: 90.095 [118, 60] loss: 0.130 [118, 120] loss: 0.129 [118, 180] loss: 0.127 [118, 240] loss: 0.135 [118, 300] loss: 0.140 [118, 360] loss: 0.138 Epoch: 118 -> Loss: 0.0981592684984 Epoch: 118 -> Test Accuracy: 90.3075 [119, 60] loss: 0.130 [119, 120] loss: 0.132 [119, 180] loss: 0.130 [119, 240] loss: 0.127 [119, 300] loss: 0.137 [119, 360] loss: 0.133 Epoch: 119 -> Loss: 0.136474281549 Epoch: 119 -> Test Accuracy: 90.315 [120, 60] loss: 0.121 [120, 120] loss: 0.132 [120, 180] loss: 0.125 [120, 240] loss: 0.141 [120, 300] loss: 0.133 [120, 360] loss: 0.128 Epoch: 120 -> Loss: 0.0627753213048 Epoch: 120 -> Test Accuracy: 90.1825 [121, 60] loss: 0.096 [121, 120] loss: 0.081 [121, 180] loss: 0.072 [121, 240] loss: 0.071 [121, 300] loss: 0.069 [121, 360] loss: 0.075 Epoch: 121 -> Loss: 0.0725773051381 Epoch: 121 -> Test Accuracy: 92.2175 [122, 60] loss: 0.064 [122, 120] loss: 0.059 [122, 180] loss: 0.062 [122, 240] loss: 0.062 [122, 300] loss: 0.055 [122, 360] loss: 0.059 Epoch: 122 -> Loss: 0.0669576302171 Epoch: 122 -> Test Accuracy: 92.225 [123, 60] loss: 0.053 [123, 120] loss: 0.055 [123, 180] loss: 0.057 [123, 240] loss: 0.050 [123, 300] loss: 0.053 [123, 360] loss: 0.053 Epoch: 123 -> Loss: 0.0238902159035 Epoch: 123 -> Test Accuracy: 92.405 [124, 60] loss: 0.047 [124, 120] loss: 0.048 [124, 180] loss: 0.047 [124, 240] loss: 0.051 [124, 300] loss: 0.050 [124, 360] loss: 0.052 Epoch: 124 -> Loss: 0.0751750022173 Epoch: 124 -> Test Accuracy: 92.46 [125, 60] loss: 0.043 [125, 120] loss: 0.044 [125, 180] loss: 0.046 [125, 240] loss: 0.046 [125, 300] loss: 0.050 [125, 360] loss: 0.045 Epoch: 125 -> Loss: 0.0497423000634 Epoch: 125 -> Test Accuracy: 92.345 [126, 60] loss: 0.039 [126, 120] loss: 0.044 [126, 180] loss: 0.041 [126, 240] loss: 0.044 [126, 300] loss: 0.045 [126, 360] loss: 0.044 Epoch: 126 -> Loss: 0.0285705067217 Epoch: 126 -> Test Accuracy: 92.0975 [127, 60] loss: 0.040 [127, 120] loss: 0.041 [127, 180] loss: 0.038 [127, 240] loss: 0.040 [127, 300] loss: 0.042 [127, 360] loss: 0.044 Epoch: 127 -> Loss: 0.0293492916971 Epoch: 127 -> Test Accuracy: 92.2325 [128, 60] loss: 0.038 [128, 120] loss: 0.039 [128, 180] loss: 0.036 [128, 240] loss: 0.039 [128, 300] loss: 0.037 [128, 360] loss: 0.041 Epoch: 128 -> Loss: 0.0484538264573 Epoch: 128 -> Test Accuracy: 92.2175 [129, 60] loss: 0.038 [129, 120] loss: 0.038 [129, 180] loss: 0.038 [129, 240] loss: 0.037 [129, 300] loss: 0.038 [129, 360] loss: 0.038 Epoch: 129 -> Loss: 0.0514101460576 Epoch: 129 -> Test Accuracy: 92.2875 [130, 60] loss: 0.036 [130, 120] loss: 0.038 [130, 180] loss: 0.032 [130, 240] loss: 0.037 [130, 300] loss: 0.037 [130, 360] loss: 0.037 Epoch: 130 -> Loss: 0.0424250736833 Epoch: 130 -> Test Accuracy: 92.1325 [131, 60] loss: 0.032 [131, 120] loss: 0.034 [131, 180] loss: 0.034 [131, 240] loss: 0.033 [131, 300] loss: 0.033 [131, 360] loss: 0.033 Epoch: 131 -> Loss: 0.0391034409404 Epoch: 131 -> Test Accuracy: 92.2525 [132, 60] loss: 0.033 [132, 120] loss: 0.033 [132, 180] loss: 0.035 [132, 240] loss: 0.036 [132, 300] loss: 0.032 [132, 360] loss: 0.036 Epoch: 132 -> Loss: 0.0384169593453 Epoch: 132 -> Test Accuracy: 92.165 [133, 60] loss: 0.029 [133, 120] loss: 0.030 [133, 180] loss: 0.031 [133, 240] loss: 0.031 [133, 300] loss: 0.032 [133, 360] loss: 0.037 Epoch: 133 -> Loss: 0.014396908693 Epoch: 133 -> Test Accuracy: 92.16 [134, 60] loss: 0.033 [134, 120] loss: 0.031 [134, 180] loss: 0.032 [134, 240] loss: 0.033 [134, 300] loss: 0.033 [134, 360] loss: 0.033 Epoch: 134 -> Loss: 0.0565491244197 Epoch: 134 -> Test Accuracy: 92.0725 [135, 60] loss: 0.030 [135, 120] loss: 0.029 [135, 180] loss: 0.031 [135, 240] loss: 0.031 [135, 300] loss: 0.030 [135, 360] loss: 0.031 Epoch: 135 -> Loss: 0.0157452970743 Epoch: 135 -> Test Accuracy: 92.22 [136, 60] loss: 0.030 [136, 120] loss: 0.031 [136, 180] loss: 0.029 [136, 240] loss: 0.030 [136, 300] loss: 0.030 [136, 360] loss: 0.029 Epoch: 136 -> Loss: 0.0522452518344 Epoch: 136 -> Test Accuracy: 92.26 [137, 60] loss: 0.028 [137, 120] loss: 0.027 [137, 180] loss: 0.029 [137, 240] loss: 0.031 [137, 300] loss: 0.031 [137, 360] loss: 0.032 Epoch: 137 -> Loss: 0.0168153364211 Epoch: 137 -> Test Accuracy: 92.0925 [138, 60] loss: 0.026 [138, 120] loss: 0.025 [138, 180] loss: 0.028 [138, 240] loss: 0.032 [138, 300] loss: 0.029 [138, 360] loss: 0.029 Epoch: 138 -> Loss: 0.0144899655133 Epoch: 138 -> Test Accuracy: 92.1325 [139, 60] loss: 0.030 [139, 120] loss: 0.030 [139, 180] loss: 0.028 [139, 240] loss: 0.029 [139, 300] loss: 0.029 [139, 360] loss: 0.029 Epoch: 139 -> Loss: 0.0294731017202 Epoch: 139 -> Test Accuracy: 92.0575 [140, 60] loss: 0.028 [140, 120] loss: 0.027 [140, 180] loss: 0.025 [140, 240] loss: 0.030 [140, 300] loss: 0.029 [140, 360] loss: 0.030 Epoch: 140 -> Loss: 0.0386446416378 Epoch: 140 -> Test Accuracy: 92.1675 [141, 60] loss: 0.026 [141, 120] loss: 0.026 [141, 180] loss: 0.027 [141, 240] loss: 0.029 [141, 300] loss: 0.025 [141, 360] loss: 0.026 Epoch: 141 -> Loss: 0.0462640114129 Epoch: 141 -> Test Accuracy: 92.0175 [142, 60] loss: 0.024 [142, 120] loss: 0.026 [142, 180] loss: 0.023 [142, 240] loss: 0.026 [142, 300] loss: 0.028 [142, 360] loss: 0.027 Epoch: 142 -> Loss: 0.0198109000921 Epoch: 142 -> Test Accuracy: 92.19 [143, 60] loss: 0.025 [143, 120] loss: 0.026 [143, 180] loss: 0.029 [143, 240] loss: 0.028 [143, 300] loss: 0.026 [143, 360] loss: 0.027 Epoch: 143 -> Loss: 0.0327403433621 Epoch: 143 -> Test Accuracy: 92.03 [144, 60] loss: 0.024 [144, 120] loss: 0.026 [144, 180] loss: 0.027 [144, 240] loss: 0.025 [144, 300] loss: 0.027 [144, 360] loss: 0.027 Epoch: 144 -> Loss: 0.0148653211072 Epoch: 144 -> Test Accuracy: 91.95 [145, 60] loss: 0.024 [145, 120] loss: 0.023 [145, 180] loss: 0.024 [145, 240] loss: 0.026 [145, 300] loss: 0.024 [145, 360] loss: 0.028 Epoch: 145 -> Loss: 0.0257855504751 Epoch: 145 -> Test Accuracy: 92.075 [146, 60] loss: 0.025 [146, 120] loss: 0.025 [146, 180] loss: 0.024 [146, 240] loss: 0.025 [146, 300] loss: 0.026 [146, 360] loss: 0.024 Epoch: 146 -> Loss: 0.0358236059546 Epoch: 146 -> Test Accuracy: 92.0125 [147, 60] loss: 0.025 [147, 120] loss: 0.023 [147, 180] loss: 0.025 [147, 240] loss: 0.026 [147, 300] loss: 0.026 [147, 360] loss: 0.028 Epoch: 147 -> Loss: 0.0306266844273 Epoch: 147 -> Test Accuracy: 91.9525 [148, 60] loss: 0.023 [148, 120] loss: 0.023 [148, 180] loss: 0.025 [148, 240] loss: 0.023 [148, 300] loss: 0.029 [148, 360] loss: 0.027 Epoch: 148 -> Loss: 0.0236614309251 Epoch: 148 -> Test Accuracy: 92.1275 [149, 60] loss: 0.025 [149, 120] loss: 0.023 [149, 180] loss: 0.024 [149, 240] loss: 0.025 [149, 300] loss: 0.028 [149, 360] loss: 0.027 Epoch: 149 -> Loss: 0.013561449945 Epoch: 149 -> Test Accuracy: 91.935 [150, 60] loss: 0.024 [150, 120] loss: 0.024 [150, 180] loss: 0.025 [150, 240] loss: 0.025 [150, 300] loss: 0.026 [150, 360] loss: 0.025 Epoch: 150 -> Loss: 0.0366351529956 Epoch: 150 -> Test Accuracy: 92.0575 [151, 60] loss: 0.022 [151, 120] loss: 0.024 [151, 180] loss: 0.023 [151, 240] loss: 0.025 [151, 300] loss: 0.025 [151, 360] loss: 0.027 Epoch: 151 -> Loss: 0.0221746247262 Epoch: 151 -> Test Accuracy: 92.03 [152, 60] loss: 0.022 [152, 120] loss: 0.024 [152, 180] loss: 0.025 [152, 240] loss: 0.022 [152, 300] loss: 0.026 [152, 360] loss: 0.024 Epoch: 152 -> Loss: 0.0234043058008 Epoch: 152 -> Test Accuracy: 91.7625 [153, 60] loss: 0.023 [153, 120] loss: 0.022 [153, 180] loss: 0.023 [153, 240] loss: 0.025 [153, 300] loss: 0.023 [153, 360] loss: 0.026 Epoch: 153 -> Loss: 0.0341323800385 Epoch: 153 -> Test Accuracy: 91.845 [154, 60] loss: 0.023 [154, 120] loss: 0.023 [154, 180] loss: 0.024 [154, 240] loss: 0.025 [154, 300] loss: 0.024 [154, 360] loss: 0.024 Epoch: 154 -> Loss: 0.0257519632578 Epoch: 154 -> Test Accuracy: 91.755 [155, 60] loss: 0.022 [155, 120] loss: 0.023 [155, 180] loss: 0.026 [155, 240] loss: 0.024 [155, 300] loss: 0.025 [155, 360] loss: 0.026 Epoch: 155 -> Loss: 0.0171002503484 Epoch: 155 -> Test Accuracy: 91.93 [156, 60] loss: 0.021 [156, 120] loss: 0.021 [156, 180] loss: 0.024 [156, 240] loss: 0.024 [156, 300] loss: 0.023 [156, 360] loss: 0.026 Epoch: 156 -> Loss: 0.0187948737293 Epoch: 156 -> Test Accuracy: 91.6375 [157, 60] loss: 0.021 [157, 120] loss: 0.022 [157, 180] loss: 0.023 [157, 240] loss: 0.022 [157, 300] loss: 0.025 [157, 360] loss: 0.025 Epoch: 157 -> Loss: 0.0213442686945 Epoch: 157 -> Test Accuracy: 91.82 [158, 60] loss: 0.025 [158, 120] loss: 0.025 [158, 180] loss: 0.025 [158, 240] loss: 0.026 [158, 300] loss: 0.025 [158, 360] loss: 0.028 Epoch: 158 -> Loss: 0.0364259406924 Epoch: 158 -> Test Accuracy: 92.09 [159, 60] loss: 0.023 [159, 120] loss: 0.025 [159, 180] loss: 0.024 [159, 240] loss: 0.023 [159, 300] loss: 0.024 [159, 360] loss: 0.026 Epoch: 159 -> Loss: 0.0178028680384 Epoch: 159 -> Test Accuracy: 91.6875 [160, 60] loss: 0.024 [160, 120] loss: 0.023 [160, 180] loss: 0.023 [160, 240] loss: 0.025 [160, 300] loss: 0.027 [160, 360] loss: 0.026 Epoch: 160 -> Loss: 0.0754204690456 Epoch: 160 -> Test Accuracy: 91.7 [161, 60] loss: 0.021 [161, 120] loss: 0.019 [161, 180] loss: 0.019 [161, 240] loss: 0.017 [161, 300] loss: 0.017 [161, 360] loss: 0.016 Epoch: 161 -> Loss: 0.0247515030205 Epoch: 161 -> Test Accuracy: 92.1075 [162, 60] loss: 0.013 [162, 120] loss: 0.015 [162, 180] loss: 0.016 [162, 240] loss: 0.015 [162, 300] loss: 0.015 [162, 360] loss: 0.014 Epoch: 162 -> Loss: 0.0142389312387 Epoch: 162 -> Test Accuracy: 92.19 [163, 60] loss: 0.014 [163, 120] loss: 0.014 [163, 180] loss: 0.014 [163, 240] loss: 0.014 [163, 300] loss: 0.015 [163, 360] loss: 0.013 Epoch: 163 -> Loss: 0.0101729640737 Epoch: 163 -> Test Accuracy: 92.195 [164, 60] loss: 0.013 [164, 120] loss: 0.013 [164, 180] loss: 0.013 [164, 240] loss: 0.013 [164, 300] loss: 0.013 [164, 360] loss: 0.014 Epoch: 164 -> Loss: 0.0181327145547 Epoch: 164 -> Test Accuracy: 92.16 [165, 60] loss: 0.013 [165, 120] loss: 0.013 [165, 180] loss: 0.013 [165, 240] loss: 0.012 [165, 300] loss: 0.012 [165, 360] loss: 0.013 Epoch: 165 -> Loss: 0.00228249793872 Epoch: 165 -> Test Accuracy: 92.2325 [166, 60] loss: 0.013 [166, 120] loss: 0.012 [166, 180] loss: 0.012 [166, 240] loss: 0.013 [166, 300] loss: 0.013 [166, 360] loss: 0.012 Epoch: 166 -> Loss: 0.00735792284831 Epoch: 166 -> Test Accuracy: 92.18 [167, 60] loss: 0.012 [167, 120] loss: 0.011 [167, 180] loss: 0.013 [167, 240] loss: 0.012 [167, 300] loss: 0.013 [167, 360] loss: 0.012 Epoch: 167 -> Loss: 0.00716293137521 Epoch: 167 -> Test Accuracy: 92.15 [168, 60] loss: 0.012 [168, 120] loss: 0.010 [168, 180] loss: 0.012 [168, 240] loss: 0.012 [168, 300] loss: 0.012 [168, 360] loss: 0.013 Epoch: 168 -> Loss: 0.0218465216458 Epoch: 168 -> Test Accuracy: 92.205 [169, 60] loss: 0.012 [169, 120] loss: 0.010 [169, 180] loss: 0.011 [169, 240] loss: 0.012 [169, 300] loss: 0.011 [169, 360] loss: 0.011 Epoch: 169 -> Loss: 0.0141726005822 Epoch: 169 -> Test Accuracy: 92.3075 [170, 60] loss: 0.012 [170, 120] loss: 0.012 [170, 180] loss: 0.012 [170, 240] loss: 0.011 [170, 300] loss: 0.012 [170, 360] loss: 0.011 Epoch: 170 -> Loss: 0.0133976582438 Epoch: 170 -> Test Accuracy: 92.21 [171, 60] loss: 0.011 [171, 120] loss: 0.011 [171, 180] loss: 0.012 [171, 240] loss: 0.011 [171, 300] loss: 0.011 [171, 360] loss: 0.012 Epoch: 171 -> Loss: 0.0172377824783 Epoch: 171 -> Test Accuracy: 92.2825 [172, 60] loss: 0.010 [172, 120] loss: 0.011 [172, 180] loss: 0.011 [172, 240] loss: 0.011 [172, 300] loss: 0.011 [172, 360] loss: 0.012 Epoch: 172 -> Loss: 0.00986809376627 Epoch: 172 -> Test Accuracy: 92.2425 [173, 60] loss: 0.010 [173, 120] loss: 0.010 [173, 180] loss: 0.011 [173, 240] loss: 0.012 [173, 300] loss: 0.010 [173, 360] loss: 0.011 Epoch: 173 -> Loss: 0.0149984331802 Epoch: 173 -> Test Accuracy: 92.2925 [174, 60] loss: 0.011 [174, 120] loss: 0.011 [174, 180] loss: 0.011 [174, 240] loss: 0.011 [174, 300] loss: 0.010 [174, 360] loss: 0.011 Epoch: 174 -> Loss: 0.00998050533235 Epoch: 174 -> Test Accuracy: 92.2625 [175, 60] loss: 0.010 [175, 120] loss: 0.011 [175, 180] loss: 0.012 [175, 240] loss: 0.011 [175, 300] loss: 0.010 [175, 360] loss: 0.010 Epoch: 175 -> Loss: 0.0101813357323 Epoch: 175 -> Test Accuracy: 92.215 [176, 60] loss: 0.010 [176, 120] loss: 0.010 [176, 180] loss: 0.011 [176, 240] loss: 0.011 [176, 300] loss: 0.010 [176, 360] loss: 0.011 Epoch: 176 -> Loss: 0.011940584518 Epoch: 176 -> Test Accuracy: 92.2125 [177, 60] loss: 0.010 [177, 120] loss: 0.011 [177, 180] loss: 0.010 [177, 240] loss: 0.011 [177, 300] loss: 0.010 [177, 360] loss: 0.010 Epoch: 177 -> Loss: 0.00595696875826 Epoch: 177 -> Test Accuracy: 92.25 [178, 60] loss: 0.011 [178, 120] loss: 0.011 [178, 180] loss: 0.010 [178, 240] loss: 0.010 [178, 300] loss: 0.010 [178, 360] loss: 0.010 Epoch: 178 -> Loss: 0.00822608359158 Epoch: 178 -> Test Accuracy: 92.2025 [179, 60] loss: 0.010 [179, 120] loss: 0.010 [179, 180] loss: 0.010 [179, 240] loss: 0.011 [179, 300] loss: 0.010 [179, 360] loss: 0.010 Epoch: 179 -> Loss: 0.00978438556194 Epoch: 179 -> Test Accuracy: 92.21 [180, 60] loss: 0.010 [180, 120] loss: 0.010 [180, 180] loss: 0.010 [180, 240] loss: 0.010 [180, 300] loss: 0.011 [180, 360] loss: 0.010 Epoch: 180 -> Loss: 0.00779106328264 Epoch: 180 -> Test Accuracy: 92.2375 [181, 60] loss: 0.010 [181, 120] loss: 0.010 [181, 180] loss: 0.011 [181, 240] loss: 0.010 [181, 300] loss: 0.010 [181, 360] loss: 0.010 Epoch: 181 -> Loss: 0.00954251550138 Epoch: 181 -> Test Accuracy: 92.25 [182, 60] loss: 0.010 [182, 120] loss: 0.010 [182, 180] loss: 0.011 [182, 240] loss: 0.010 [182, 300] loss: 0.010 [182, 360] loss: 0.011 Epoch: 182 -> Loss: 0.00503360107541 Epoch: 182 -> Test Accuracy: 92.2475 [183, 60] loss: 0.011 [183, 120] loss: 0.010 [183, 180] loss: 0.009 [183, 240] loss: 0.010 [183, 300] loss: 0.011 [183, 360] loss: 0.011 Epoch: 183 -> Loss: 0.0226580798626 Epoch: 183 -> Test Accuracy: 92.21 [184, 60] loss: 0.009 [184, 120] loss: 0.010 [184, 180] loss: 0.010 [184, 240] loss: 0.010 [184, 300] loss: 0.009 [184, 360] loss: 0.010 Epoch: 184 -> Loss: 0.0174920354038 Epoch: 184 -> Test Accuracy: 92.275 [185, 60] loss: 0.010 [185, 120] loss: 0.009 [185, 180] loss: 0.010 [185, 240] loss: 0.010 [185, 300] loss: 0.010 [185, 360] loss: 0.010 Epoch: 185 -> Loss: 0.00887467432767 Epoch: 185 -> Test Accuracy: 92.2 [186, 60] loss: 0.009 [186, 120] loss: 0.009 [186, 180] loss: 0.010 [186, 240] loss: 0.010 [186, 300] loss: 0.010 [186, 360] loss: 0.011 Epoch: 186 -> Loss: 0.0100764958188 Epoch: 186 -> Test Accuracy: 92.2775 [187, 60] loss: 0.009 [187, 120] loss: 0.009 [187, 180] loss: 0.010 [187, 240] loss: 0.010 [187, 300] loss: 0.010 [187, 360] loss: 0.010 Epoch: 187 -> Loss: 0.00767189497128 Epoch: 187 -> Test Accuracy: 92.2925 [188, 60] loss: 0.009 [188, 120] loss: 0.011 [188, 180] loss: 0.009 [188, 240] loss: 0.009 [188, 300] loss: 0.009 [188, 360] loss: 0.010 Epoch: 188 -> Loss: 0.0170344524086 Epoch: 188 -> Test Accuracy: 92.23 [189, 60] loss: 0.009 [189, 120] loss: 0.009 [189, 180] loss: 0.010 [189, 240] loss: 0.010 [189, 300] loss: 0.010 [189, 360] loss: 0.010 Epoch: 189 -> Loss: 0.0175261907279 Epoch: 189 -> Test Accuracy: 92.2 [190, 60] loss: 0.010 [190, 120] loss: 0.010 [190, 180] loss: 0.009 [190, 240] loss: 0.009 [190, 300] loss: 0.009 [190, 360] loss: 0.010 Epoch: 190 -> Loss: 0.0128434095532 Epoch: 190 -> Test Accuracy: 92.21 [191, 60] loss: 0.010 [191, 120] loss: 0.010 [191, 180] loss: 0.009 [191, 240] loss: 0.010 [191, 300] loss: 0.010 [191, 360] loss: 0.009 Epoch: 191 -> Loss: 0.0104228034616 Epoch: 191 -> Test Accuracy: 92.185 [192, 60] loss: 0.010 [192, 120] loss: 0.010 [192, 180] loss: 0.010 [192, 240] loss: 0.010 [192, 300] loss: 0.010 [192, 360] loss: 0.009 Epoch: 192 -> Loss: 0.00403060019016 Epoch: 192 -> Test Accuracy: 92.1775 [193, 60] loss: 0.009 [193, 120] loss: 0.010 [193, 180] loss: 0.010 [193, 240] loss: 0.009 [193, 300] loss: 0.009 [193, 360] loss: 0.009 Epoch: 193 -> Loss: 0.0098797082901 Epoch: 193 -> Test Accuracy: 92.1575 [194, 60] loss: 0.009 [194, 120] loss: 0.009 [194, 180] loss: 0.010 [194, 240] loss: 0.010 [194, 300] loss: 0.009 [194, 360] loss: 0.010 Epoch: 194 -> Loss: 0.00674241501838 Epoch: 194 -> Test Accuracy: 92.1775 [195, 60] loss: 0.010 [195, 120] loss: 0.009 [195, 180] loss: 0.009 [195, 240] loss: 0.010 [195, 300] loss: 0.009 [195, 360] loss: 0.009 Epoch: 195 -> Loss: 0.00723540503532 Epoch: 195 -> Test Accuracy: 92.145 [196, 60] loss: 0.009 [196, 120] loss: 0.009 [196, 180] loss: 0.009 [196, 240] loss: 0.009 [196, 300] loss: 0.010 [196, 360] loss: 0.010 Epoch: 196 -> Loss: 0.00891743786633 Epoch: 196 -> Test Accuracy: 92.1725 [197, 60] loss: 0.009 [197, 120] loss: 0.009 [197, 180] loss: 0.009 [197, 240] loss: 0.010 [197, 300] loss: 0.009 [197, 360] loss: 0.010 Epoch: 197 -> Loss: 0.0150486528873 Epoch: 197 -> Test Accuracy: 92.1625 [198, 60] loss: 0.008 [198, 120] loss: 0.009 [198, 180] loss: 0.009 [198, 240] loss: 0.010 [198, 300] loss: 0.009 [198, 360] loss: 0.009 Epoch: 198 -> Loss: 0.00717498268932 Epoch: 198 -> Test Accuracy: 92.175 [199, 60] loss: 0.009 [199, 120] loss: 0.009 [199, 180] loss: 0.009 [199, 240] loss: 0.010 [199, 300] loss: 0.009 [199, 360] loss: 0.010 Epoch: 199 -> Loss: 0.00609672721475 Epoch: 199 -> Test Accuracy: 92.165 [200, 60] loss: 0.009 [200, 120] loss: 0.009 [200, 180] loss: 0.009 [200, 240] loss: 0.010 [200, 300] loss: 0.009 [200, 360] loss: 0.009 Epoch: 200 -> Loss: 0.00567238777876 Epoch: 200 -> Test Accuracy: 92.19 Finished Training
In [14]:
# train NonLinearClassifiers on feature map of net_3block
block3_loss_log, _, block3_test_accuracy_log, _, _ = tr.train_all_blocks(3, 10, [0.1, 0.02, 0.004, 0.0008],
[20, 40, 45, 100], 0.9, 5e-4, net_block3, criterion, trainloader, None, testloader)
[1, 60] loss: 2.135 [1, 120] loss: 1.237 [1, 180] loss: 1.118 [1, 240] loss: 1.054 [1, 300] loss: 1.027 [1, 360] loss: 0.987 Epoch: 1 -> Loss: 1.02406382561 Epoch: 1 -> Test Accuracy: 68.73 [2, 60] loss: 0.908 [2, 120] loss: 0.900 [2, 180] loss: 0.892 [2, 240] loss: 0.870 [2, 300] loss: 0.870 [2, 360] loss: 0.830 Epoch: 2 -> Loss: 0.819397449493 Epoch: 2 -> Test Accuracy: 72.48 [3, 60] loss: 0.808 [3, 120] loss: 0.809 [3, 180] loss: 0.789 [3, 240] loss: 0.807 [3, 300] loss: 0.777 [3, 360] loss: 0.760 Epoch: 3 -> Loss: 0.705734670162 Epoch: 3 -> Test Accuracy: 74.34 [4, 60] loss: 0.757 [4, 120] loss: 0.763 [4, 180] loss: 0.744 [4, 240] loss: 0.729 [4, 300] loss: 0.732 [4, 360] loss: 0.722 Epoch: 4 -> Loss: 0.572060704231 Epoch: 4 -> Test Accuracy: 75.96 [5, 60] loss: 0.705 [5, 120] loss: 0.700 [5, 180] loss: 0.710 [5, 240] loss: 0.693 [5, 300] loss: 0.706 [5, 360] loss: 0.720 Epoch: 5 -> Loss: 0.530122101307 Epoch: 5 -> Test Accuracy: 76.62 [6, 60] loss: 0.681 [6, 120] loss: 0.698 [6, 180] loss: 0.682 [6, 240] loss: 0.688 [6, 300] loss: 0.689 [6, 360] loss: 0.684 Epoch: 6 -> Loss: 0.762083053589 Epoch: 6 -> Test Accuracy: 77.13 [7, 60] loss: 0.647 [7, 120] loss: 0.658 [7, 180] loss: 0.682 [7, 240] loss: 0.670 [7, 300] loss: 0.683 [7, 360] loss: 0.668 Epoch: 7 -> Loss: 0.561828672886 Epoch: 7 -> Test Accuracy: 76.95 [8, 60] loss: 0.643 [8, 120] loss: 0.646 [8, 180] loss: 0.652 [8, 240] loss: 0.667 [8, 300] loss: 0.651 [8, 360] loss: 0.660 Epoch: 8 -> Loss: 0.631638646126 Epoch: 8 -> Test Accuracy: 77.92 [9, 60] loss: 0.617 [9, 120] loss: 0.627 [9, 180] loss: 0.632 [9, 240] loss: 0.663 [9, 300] loss: 0.652 [9, 360] loss: 0.653 Epoch: 9 -> Loss: 0.601558744907 Epoch: 9 -> Test Accuracy: 77.99 [10, 60] loss: 0.607 [10, 120] loss: 0.628 [10, 180] loss: 0.627 [10, 240] loss: 0.632 [10, 300] loss: 0.629 [10, 360] loss: 0.639 Epoch: 10 -> Loss: 0.69752061367 Epoch: 10 -> Test Accuracy: 77.83 [11, 60] loss: 0.615 [11, 120] loss: 0.618 [11, 180] loss: 0.613 [11, 240] loss: 0.626 [11, 300] loss: 0.623 [11, 360] loss: 0.626 Epoch: 11 -> Loss: 0.627347946167 Epoch: 11 -> Test Accuracy: 78.23 [12, 60] loss: 0.615 [12, 120] loss: 0.603 [12, 180] loss: 0.591 [12, 240] loss: 0.627 [12, 300] loss: 0.607 [12, 360] loss: 0.641 Epoch: 12 -> Loss: 0.440225422382 Epoch: 12 -> Test Accuracy: 78.78 [13, 60] loss: 0.601 [13, 120] loss: 0.604 [13, 180] loss: 0.606 [13, 240] loss: 0.617 [13, 300] loss: 0.617 [13, 360] loss: 0.624 Epoch: 13 -> Loss: 0.530610501766 Epoch: 13 -> Test Accuracy: 78.81 [14, 60] loss: 0.568 [14, 120] loss: 0.587 [14, 180] loss: 0.621 [14, 240] loss: 0.614 [14, 300] loss: 0.617 [14, 360] loss: 0.612 Epoch: 14 -> Loss: 0.864044070244 Epoch: 14 -> Test Accuracy: 78.74 [15, 60] loss: 0.597 [15, 120] loss: 0.594 [15, 180] loss: 0.586 [15, 240] loss: 0.598 [15, 300] loss: 0.578 [15, 360] loss: 0.624 Epoch: 15 -> Loss: 0.784435331821 Epoch: 15 -> Test Accuracy: 78.56 [16, 60] loss: 0.594 [16, 120] loss: 0.591 [16, 180] loss: 0.584 [16, 240] loss: 0.606 [16, 300] loss: 0.626 [16, 360] loss: 0.591 Epoch: 16 -> Loss: 0.565725684166 Epoch: 16 -> Test Accuracy: 78.61 [17, 60] loss: 0.573 [17, 120] loss: 0.581 [17, 180] loss: 0.611 [17, 240] loss: 0.597 [17, 300] loss: 0.590 [17, 360] loss: 0.621 Epoch: 17 -> Loss: 0.595977246761 Epoch: 17 -> Test Accuracy: 78.18 [18, 60] loss: 0.570 [18, 120] loss: 0.588 [18, 180] loss: 0.570 [18, 240] loss: 0.595 [18, 300] loss: 0.614 [18, 360] loss: 0.604 Epoch: 18 -> Loss: 0.599996745586 Epoch: 18 -> Test Accuracy: 78.87 [19, 60] loss: 0.558 [19, 120] loss: 0.578 [19, 180] loss: 0.579 [19, 240] loss: 0.604 [19, 300] loss: 0.611 [19, 360] loss: 0.589 Epoch: 19 -> Loss: 0.615851283073 Epoch: 19 -> Test Accuracy: 78.81 [20, 60] loss: 0.558 [20, 120] loss: 0.579 [20, 180] loss: 0.583 [20, 240] loss: 0.590 [20, 300] loss: 0.596 [20, 360] loss: 0.610 Epoch: 20 -> Loss: 0.605484127998 Epoch: 20 -> Test Accuracy: 79.39 [21, 60] loss: 0.538 [21, 120] loss: 0.491 [21, 180] loss: 0.480 [21, 240] loss: 0.500 [21, 300] loss: 0.482 [21, 360] loss: 0.483 Epoch: 21 -> Loss: 0.56816971302 Epoch: 21 -> Test Accuracy: 81.5 [22, 60] loss: 0.441 [22, 120] loss: 0.454 [22, 180] loss: 0.459 [22, 240] loss: 0.450 [22, 300] loss: 0.442 [22, 360] loss: 0.451 Epoch: 22 -> Loss: 0.40216255188 Epoch: 22 -> Test Accuracy: 81.47 [23, 60] loss: 0.421 [23, 120] loss: 0.454 [23, 180] loss: 0.422 [23, 240] loss: 0.433 [23, 300] loss: 0.426 [23, 360] loss: 0.438 Epoch: 23 -> Loss: 0.486105382442 Epoch: 23 -> Test Accuracy: 82.19 [24, 60] loss: 0.420 [24, 120] loss: 0.432 [24, 180] loss: 0.421 [24, 240] loss: 0.428 [24, 300] loss: 0.409 [24, 360] loss: 0.417 Epoch: 24 -> Loss: 0.365364342928 Epoch: 24 -> Test Accuracy: 82.21 [25, 60] loss: 0.403 [25, 120] loss: 0.411 [25, 180] loss: 0.400 [25, 240] loss: 0.418 [25, 300] loss: 0.423 [25, 360] loss: 0.421 Epoch: 25 -> Loss: 0.491146504879 Epoch: 25 -> Test Accuracy: 82.11 [26, 60] loss: 0.414 [26, 120] loss: 0.407 [26, 180] loss: 0.422 [26, 240] loss: 0.406 [26, 300] loss: 0.424 [26, 360] loss: 0.397 Epoch: 26 -> Loss: 0.56111395359 Epoch: 26 -> Test Accuracy: 82.16 [27, 60] loss: 0.391 [27, 120] loss: 0.405 [27, 180] loss: 0.395 [27, 240] loss: 0.395 [27, 300] loss: 0.404 [27, 360] loss: 0.401 Epoch: 27 -> Loss: 0.438163995743 Epoch: 27 -> Test Accuracy: 82.31 [28, 60] loss: 0.393 [28, 120] loss: 0.385 [28, 180] loss: 0.393 [28, 240] loss: 0.382 [28, 300] loss: 0.408 [28, 360] loss: 0.407 Epoch: 28 -> Loss: 0.282507956028 Epoch: 28 -> Test Accuracy: 82.44 [29, 60] loss: 0.390 [29, 120] loss: 0.374 [29, 180] loss: 0.378 [29, 240] loss: 0.403 [29, 300] loss: 0.394 [29, 360] loss: 0.404 Epoch: 29 -> Loss: 0.281484305859 Epoch: 29 -> Test Accuracy: 82.13 [30, 60] loss: 0.381 [30, 120] loss: 0.370 [30, 180] loss: 0.401 [30, 240] loss: 0.381 [30, 300] loss: 0.384 [30, 360] loss: 0.404 Epoch: 30 -> Loss: 0.277964651585 Epoch: 30 -> Test Accuracy: 81.73 [31, 60] loss: 0.382 [31, 120] loss: 0.378 [31, 180] loss: 0.388 [31, 240] loss: 0.387 [31, 300] loss: 0.392 [31, 360] loss: 0.378 Epoch: 31 -> Loss: 0.347091972828 Epoch: 31 -> Test Accuracy: 82.47 [32, 60] loss: 0.390 [32, 120] loss: 0.366 [32, 180] loss: 0.377 [32, 240] loss: 0.386 [32, 300] loss: 0.382 [32, 360] loss: 0.396 Epoch: 32 -> Loss: 0.42511588335 Epoch: 32 -> Test Accuracy: 81.95 [33, 60] loss: 0.371 [33, 120] loss: 0.368 [33, 180] loss: 0.377 [33, 240] loss: 0.384 [33, 300] loss: 0.388 [33, 360] loss: 0.378 Epoch: 33 -> Loss: 0.391119420528 Epoch: 33 -> Test Accuracy: 81.62 [34, 60] loss: 0.377 [34, 120] loss: 0.355 [34, 180] loss: 0.373 [34, 240] loss: 0.394 [34, 300] loss: 0.390 [34, 360] loss: 0.394 Epoch: 34 -> Loss: 0.403971672058 Epoch: 34 -> Test Accuracy: 82.03 [35, 60] loss: 0.379 [35, 120] loss: 0.371 [35, 180] loss: 0.377 [35, 240] loss: 0.385 [35, 300] loss: 0.393 [35, 360] loss: 0.386 Epoch: 35 -> Loss: 0.37564048171 Epoch: 35 -> Test Accuracy: 82.11 [36, 60] loss: 0.372 [36, 120] loss: 0.369 [36, 180] loss: 0.385 [36, 240] loss: 0.383 [36, 300] loss: 0.367 [36, 360] loss: 0.384 Epoch: 36 -> Loss: 0.331483066082 Epoch: 36 -> Test Accuracy: 81.55 [37, 60] loss: 0.365 [37, 120] loss: 0.376 [37, 180] loss: 0.365 [37, 240] loss: 0.372 [37, 300] loss: 0.384 [37, 360] loss: 0.388 Epoch: 37 -> Loss: 0.326968252659 Epoch: 37 -> Test Accuracy: 82.09 [38, 60] loss: 0.372 [38, 120] loss: 0.373 [38, 180] loss: 0.386 [38, 240] loss: 0.368 [38, 300] loss: 0.375 [38, 360] loss: 0.385 Epoch: 38 -> Loss: 0.287506878376 Epoch: 38 -> Test Accuracy: 82.01 [39, 60] loss: 0.361 [39, 120] loss: 0.375 [39, 180] loss: 0.358 [39, 240] loss: 0.384 [39, 300] loss: 0.375 [39, 360] loss: 0.362 Epoch: 39 -> Loss: 0.385202825069 Epoch: 39 -> Test Accuracy: 81.53 [40, 60] loss: 0.348 [40, 120] loss: 0.369 [40, 180] loss: 0.376 [40, 240] loss: 0.391 [40, 300] loss: 0.361 [40, 360] loss: 0.381 Epoch: 40 -> Loss: 0.411934942007 Epoch: 40 -> Test Accuracy: 82.03 [41, 60] loss: 0.352 [41, 120] loss: 0.322 [41, 180] loss: 0.322 [41, 240] loss: 0.336 [41, 300] loss: 0.333 [41, 360] loss: 0.323 Epoch: 41 -> Loss: 0.234177619219 Epoch: 41 -> Test Accuracy: 82.78 [42, 60] loss: 0.307 [42, 120] loss: 0.310 [42, 180] loss: 0.319 [42, 240] loss: 0.307 [42, 300] loss: 0.315 [42, 360] loss: 0.308 Epoch: 42 -> Loss: 0.332994103432 Epoch: 42 -> Test Accuracy: 82.73 [43, 60] loss: 0.294 [43, 120] loss: 0.293 [43, 180] loss: 0.297 [43, 240] loss: 0.304 [43, 300] loss: 0.298 [43, 360] loss: 0.299 Epoch: 43 -> Loss: 0.326020866632 Epoch: 43 -> Test Accuracy: 83.09 [44, 60] loss: 0.295 [44, 120] loss: 0.292 [44, 180] loss: 0.283 [44, 240] loss: 0.294 [44, 300] loss: 0.293 [44, 360] loss: 0.282 Epoch: 44 -> Loss: 0.404775053263 Epoch: 44 -> Test Accuracy: 82.88 [45, 60] loss: 0.272 [45, 120] loss: 0.282 [45, 180] loss: 0.285 [45, 240] loss: 0.288 [45, 300] loss: 0.278 [45, 360] loss: 0.299 Epoch: 45 -> Loss: 0.15308393538 Epoch: 45 -> Test Accuracy: 82.92 [46, 60] loss: 0.275 [46, 120] loss: 0.268 [46, 180] loss: 0.283 [46, 240] loss: 0.278 [46, 300] loss: 0.278 [46, 360] loss: 0.270 Epoch: 46 -> Loss: 0.262244164944 Epoch: 46 -> Test Accuracy: 83.05 [47, 60] loss: 0.276 [47, 120] loss: 0.267 [47, 180] loss: 0.275 [47, 240] loss: 0.279 [47, 300] loss: 0.266 [47, 360] loss: 0.262 Epoch: 47 -> Loss: 0.361472249031 Epoch: 47 -> Test Accuracy: 83.08 [48, 60] loss: 0.270 [48, 120] loss: 0.271 [48, 180] loss: 0.268 [48, 240] loss: 0.265 [48, 300] loss: 0.285 [48, 360] loss: 0.286 Epoch: 48 -> Loss: 0.355505049229 Epoch: 48 -> Test Accuracy: 83.16 [49, 60] loss: 0.272 [49, 120] loss: 0.270 [49, 180] loss: 0.271 [49, 240] loss: 0.268 [49, 300] loss: 0.257 [49, 360] loss: 0.266 Epoch: 49 -> Loss: 0.208222582936 Epoch: 49 -> Test Accuracy: 83.07 [50, 60] loss: 0.264 [50, 120] loss: 0.266 [50, 180] loss: 0.266 [50, 240] loss: 0.277 [50, 300] loss: 0.265 [50, 360] loss: 0.255 Epoch: 50 -> Loss: 0.365976423025 Epoch: 50 -> Test Accuracy: 82.99 [51, 60] loss: 0.264 [51, 120] loss: 0.262 [51, 180] loss: 0.250 [51, 240] loss: 0.259 [51, 300] loss: 0.263 [51, 360] loss: 0.264 Epoch: 51 -> Loss: 0.314839750528 Epoch: 51 -> Test Accuracy: 82.98 [52, 60] loss: 0.262 [52, 120] loss: 0.259 [52, 180] loss: 0.260 [52, 240] loss: 0.252 [52, 300] loss: 0.252 [52, 360] loss: 0.275 Epoch: 52 -> Loss: 0.206320613623 Epoch: 52 -> Test Accuracy: 83.04 [53, 60] loss: 0.267 [53, 120] loss: 0.260 [53, 180] loss: 0.267 [53, 240] loss: 0.255 [53, 300] loss: 0.267 [53, 360] loss: 0.263 Epoch: 53 -> Loss: 0.397431999445 Epoch: 53 -> Test Accuracy: 82.94 [54, 60] loss: 0.250 [54, 120] loss: 0.260 [54, 180] loss: 0.260 [54, 240] loss: 0.259 [54, 300] loss: 0.261 [54, 360] loss: 0.251 Epoch: 54 -> Loss: 0.355315774679 Epoch: 54 -> Test Accuracy: 83.27 [55, 60] loss: 0.255 [55, 120] loss: 0.270 [55, 180] loss: 0.253 [55, 240] loss: 0.236 [55, 300] loss: 0.264 [55, 360] loss: 0.265 Epoch: 55 -> Loss: 0.358705937862 Epoch: 55 -> Test Accuracy: 83.2 [56, 60] loss: 0.253 [56, 120] loss: 0.256 [56, 180] loss: 0.246 [56, 240] loss: 0.245 [56, 300] loss: 0.253 [56, 360] loss: 0.262 Epoch: 56 -> Loss: 0.272285223007 Epoch: 56 -> Test Accuracy: 83.1 [57, 60] loss: 0.252 [57, 120] loss: 0.262 [57, 180] loss: 0.251 [57, 240] loss: 0.253 [57, 300] loss: 0.249 [57, 360] loss: 0.263 Epoch: 57 -> Loss: 0.406624853611 Epoch: 57 -> Test Accuracy: 83.08 [58, 60] loss: 0.256 [58, 120] loss: 0.242 [58, 180] loss: 0.259 [58, 240] loss: 0.256 [58, 300] loss: 0.251 [58, 360] loss: 0.254 Epoch: 58 -> Loss: 0.246799662709 Epoch: 58 -> Test Accuracy: 83.04 [59, 60] loss: 0.252 [59, 120] loss: 0.276 [59, 180] loss: 0.243 [59, 240] loss: 0.263 [59, 300] loss: 0.249 [59, 360] loss: 0.257 Epoch: 59 -> Loss: 0.279025554657 Epoch: 59 -> Test Accuracy: 83.19 [60, 60] loss: 0.249 [60, 120] loss: 0.249 [60, 180] loss: 0.259 [60, 240] loss: 0.247 [60, 300] loss: 0.257 [60, 360] loss: 0.252 Epoch: 60 -> Loss: 0.197952821851 Epoch: 60 -> Test Accuracy: 83.1 [61, 60] loss: 0.253 [61, 120] loss: 0.248 [61, 180] loss: 0.253 [61, 240] loss: 0.250 [61, 300] loss: 0.253 [61, 360] loss: 0.255 Epoch: 61 -> Loss: 0.305834472179 Epoch: 61 -> Test Accuracy: 83.24 [62, 60] loss: 0.244 [62, 120] loss: 0.255 [62, 180] loss: 0.263 [62, 240] loss: 0.252 [62, 300] loss: 0.246 [62, 360] loss: 0.239 Epoch: 62 -> Loss: 0.430526345968 Epoch: 62 -> Test Accuracy: 83.12 [63, 60] loss: 0.255 [63, 120] loss: 0.251 [63, 180] loss: 0.248 [63, 240] loss: 0.235 [63, 300] loss: 0.248 [63, 360] loss: 0.258 Epoch: 63 -> Loss: 0.179082587361 Epoch: 63 -> Test Accuracy: 83.29 [64, 60] loss: 0.249 [64, 120] loss: 0.246 [64, 180] loss: 0.244 [64, 240] loss: 0.243 [64, 300] loss: 0.242 [64, 360] loss: 0.241 Epoch: 64 -> Loss: 0.266464024782 Epoch: 64 -> Test Accuracy: 83.21 [65, 60] loss: 0.243 [65, 120] loss: 0.246 [65, 180] loss: 0.247 [65, 240] loss: 0.250 [65, 300] loss: 0.249 [65, 360] loss: 0.247 Epoch: 65 -> Loss: 0.241675049067 Epoch: 65 -> Test Accuracy: 83.28 [66, 60] loss: 0.229 [66, 120] loss: 0.246 [66, 180] loss: 0.255 [66, 240] loss: 0.251 [66, 300] loss: 0.235 [66, 360] loss: 0.250 Epoch: 66 -> Loss: 0.264233797789 Epoch: 66 -> Test Accuracy: 83.12 [67, 60] loss: 0.248 [67, 120] loss: 0.244 [67, 180] loss: 0.242 [67, 240] loss: 0.238 [67, 300] loss: 0.246 [67, 360] loss: 0.243 Epoch: 67 -> Loss: 0.289677709341 Epoch: 67 -> Test Accuracy: 83.22 [68, 60] loss: 0.256 [68, 120] loss: 0.247 [68, 180] loss: 0.243 [68, 240] loss: 0.241 [68, 300] loss: 0.239 [68, 360] loss: 0.234 Epoch: 68 -> Loss: 0.230375498533 Epoch: 68 -> Test Accuracy: 83.08 [69, 60] loss: 0.234 [69, 120] loss: 0.239 [69, 180] loss: 0.242 [69, 240] loss: 0.251 [69, 300] loss: 0.233 [69, 360] loss: 0.250 Epoch: 69 -> Loss: 0.277349829674 Epoch: 69 -> Test Accuracy: 83.27 [70, 60] loss: 0.244 [70, 120] loss: 0.243 [70, 180] loss: 0.244 [70, 240] loss: 0.245 [70, 300] loss: 0.246 [70, 360] loss: 0.237 Epoch: 70 -> Loss: 0.296507179737 Epoch: 70 -> Test Accuracy: 83.14 [71, 60] loss: 0.240 [71, 120] loss: 0.230 [71, 180] loss: 0.244 [71, 240] loss: 0.235 [71, 300] loss: 0.245 [71, 360] loss: 0.230 Epoch: 71 -> Loss: 0.157645359635 Epoch: 71 -> Test Accuracy: 83.04 [72, 60] loss: 0.242 [72, 120] loss: 0.229 [72, 180] loss: 0.225 [72, 240] loss: 0.238 [72, 300] loss: 0.250 [72, 360] loss: 0.234 Epoch: 72 -> Loss: 0.279527008533 Epoch: 72 -> Test Accuracy: 83.14 [73, 60] loss: 0.234 [73, 120] loss: 0.236 [73, 180] loss: 0.227 [73, 240] loss: 0.246 [73, 300] loss: 0.242 [73, 360] loss: 0.234 Epoch: 73 -> Loss: 0.251714527607 Epoch: 73 -> Test Accuracy: 83.15 [74, 60] loss: 0.231 [74, 120] loss: 0.244 [74, 180] loss: 0.247 [74, 240] loss: 0.236 [74, 300] loss: 0.242 [74, 360] loss: 0.246 Epoch: 74 -> Loss: 0.219133019447 Epoch: 74 -> Test Accuracy: 83.16 [75, 60] loss: 0.227 [75, 120] loss: 0.242 [75, 180] loss: 0.238 [75, 240] loss: 0.234 [75, 300] loss: 0.241 [75, 360] loss: 0.244 Epoch: 75 -> Loss: 0.160952612758 Epoch: 75 -> Test Accuracy: 83.21 [76, 60] loss: 0.234 [76, 120] loss: 0.231 [76, 180] loss: 0.234 [76, 240] loss: 0.241 [76, 300] loss: 0.233 [76, 360] loss: 0.233 Epoch: 76 -> Loss: 0.194017440081 Epoch: 76 -> Test Accuracy: 83.21 [77, 60] loss: 0.225 [77, 120] loss: 0.231 [77, 180] loss: 0.235 [77, 240] loss: 0.233 [77, 300] loss: 0.233 [77, 360] loss: 0.232 Epoch: 77 -> Loss: 0.230169251561 Epoch: 77 -> Test Accuracy: 83.19 [78, 60] loss: 0.231 [78, 120] loss: 0.229 [78, 180] loss: 0.233 [78, 240] loss: 0.231 [78, 300] loss: 0.236 [78, 360] loss: 0.225 Epoch: 78 -> Loss: 0.356003493071 Epoch: 78 -> Test Accuracy: 83.28 [79, 60] loss: 0.235 [79, 120] loss: 0.230 [79, 180] loss: 0.221 [79, 240] loss: 0.235 [79, 300] loss: 0.230 [79, 360] loss: 0.237 Epoch: 79 -> Loss: 0.526965022087 Epoch: 79 -> Test Accuracy: 83.32 [80, 60] loss: 0.236 [80, 120] loss: 0.234 [80, 180] loss: 0.222 [80, 240] loss: 0.233 [80, 300] loss: 0.230 [80, 360] loss: 0.227 Epoch: 80 -> Loss: 0.198395565152 Epoch: 80 -> Test Accuracy: 83.23 [81, 60] loss: 0.223 [81, 120] loss: 0.233 [81, 180] loss: 0.238 [81, 240] loss: 0.228 [81, 300] loss: 0.224 [81, 360] loss: 0.228 Epoch: 81 -> Loss: 0.368765324354 Epoch: 81 -> Test Accuracy: 83.18 [82, 60] loss: 0.227 [82, 120] loss: 0.234 [82, 180] loss: 0.230 [82, 240] loss: 0.237 [82, 300] loss: 0.225 [82, 360] loss: 0.226 Epoch: 82 -> Loss: 0.178824096918 Epoch: 82 -> Test Accuracy: 83.09 [83, 60] loss: 0.230 [83, 120] loss: 0.235 [83, 180] loss: 0.228 [83, 240] loss: 0.231 [83, 300] loss: 0.228 [83, 360] loss: 0.235 Epoch: 83 -> Loss: 0.249970510602 Epoch: 83 -> Test Accuracy: 83.29 [84, 60] loss: 0.230 [84, 120] loss: 0.223 [84, 180] loss: 0.220 [84, 240] loss: 0.232 [84, 300] loss: 0.232 [84, 360] loss: 0.223 Epoch: 84 -> Loss: 0.275760382414 Epoch: 84 -> Test Accuracy: 83.1 [85, 60] loss: 0.224 [85, 120] loss: 0.223 [85, 180] loss: 0.217 [85, 240] loss: 0.229 [85, 300] loss: 0.232 [85, 360] loss: 0.239 Epoch: 85 -> Loss: 0.303332418203 Epoch: 85 -> Test Accuracy: 83.29 [86, 60] loss: 0.220 [86, 120] loss: 0.225 [86, 180] loss: 0.224 [86, 240] loss: 0.231 [86, 300] loss: 0.228 [86, 360] loss: 0.229 Epoch: 86 -> Loss: 0.361538261175 Epoch: 86 -> Test Accuracy: 83.19 [87, 60] loss: 0.221 [87, 120] loss: 0.224 [87, 180] loss: 0.223 [87, 240] loss: 0.225 [87, 300] loss: 0.224 [87, 360] loss: 0.215 Epoch: 87 -> Loss: 0.120859101415 Epoch: 87 -> Test Accuracy: 83.2 [88, 60] loss: 0.234 [88, 120] loss: 0.215 [88, 180] loss: 0.226 [88, 240] loss: 0.217 [88, 300] loss: 0.228 [88, 360] loss: 0.231 Epoch: 88 -> Loss: 0.202250763774 Epoch: 88 -> Test Accuracy: 83.17 [89, 60] loss: 0.220 [89, 120] loss: 0.219 [89, 180] loss: 0.231 [89, 240] loss: 0.234 [89, 300] loss: 0.223 [89, 360] loss: 0.226 Epoch: 89 -> Loss: 0.243361517787 Epoch: 89 -> Test Accuracy: 83.33 [90, 60] loss: 0.227 [90, 120] loss: 0.202 [90, 180] loss: 0.225 [90, 240] loss: 0.222 [90, 300] loss: 0.221 [90, 360] loss: 0.217 Epoch: 90 -> Loss: 0.182889983058 Epoch: 90 -> Test Accuracy: 83.38 [91, 60] loss: 0.224 [91, 120] loss: 0.221 [91, 180] loss: 0.222 [91, 240] loss: 0.217 [91, 300] loss: 0.221 [91, 360] loss: 0.235 Epoch: 91 -> Loss: 0.237830594182 Epoch: 91 -> Test Accuracy: 83.29 [92, 60] loss: 0.223 [92, 120] loss: 0.225 [92, 180] loss: 0.212 [92, 240] loss: 0.226 [92, 300] loss: 0.216 [92, 360] loss: 0.229 Epoch: 92 -> Loss: 0.118292652071 Epoch: 92 -> Test Accuracy: 83.4 [93, 60] loss: 0.218 [93, 120] loss: 0.222 [93, 180] loss: 0.212 [93, 240] loss: 0.215 [93, 300] loss: 0.228 [93, 360] loss: 0.216 Epoch: 93 -> Loss: 0.167962044477 Epoch: 93 -> Test Accuracy: 83.36 [94, 60] loss: 0.222 [94, 120] loss: 0.211 [94, 180] loss: 0.213 [94, 240] loss: 0.221 [94, 300] loss: 0.217 [94, 360] loss: 0.229 Epoch: 94 -> Loss: 0.265777915716 Epoch: 94 -> Test Accuracy: 83.24 [95, 60] loss: 0.226 [95, 120] loss: 0.206 [95, 180] loss: 0.213 [95, 240] loss: 0.223 [95, 300] loss: 0.214 [95, 360] loss: 0.221 Epoch: 95 -> Loss: 0.231031134725 Epoch: 95 -> Test Accuracy: 83.35 [96, 60] loss: 0.216 [96, 120] loss: 0.219 [96, 180] loss: 0.221 [96, 240] loss: 0.214 [96, 300] loss: 0.219 [96, 360] loss: 0.219 Epoch: 96 -> Loss: 0.162611573935 Epoch: 96 -> Test Accuracy: 83.43 [97, 60] loss: 0.216 [97, 120] loss: 0.224 [97, 180] loss: 0.215 [97, 240] loss: 0.217 [97, 300] loss: 0.215 [97, 360] loss: 0.212 Epoch: 97 -> Loss: 0.159254923463 Epoch: 97 -> Test Accuracy: 83.26 [98, 60] loss: 0.206 [98, 120] loss: 0.212 [98, 180] loss: 0.214 [98, 240] loss: 0.216 [98, 300] loss: 0.223 [98, 360] loss: 0.207 Epoch: 98 -> Loss: 0.2572067976 Epoch: 98 -> Test Accuracy: 83.34 [99, 60] loss: 0.210 [99, 120] loss: 0.203 [99, 180] loss: 0.217 [99, 240] loss: 0.219 [99, 300] loss: 0.218 [99, 360] loss: 0.214 Epoch: 99 -> Loss: 0.2847032547 Epoch: 99 -> Test Accuracy: 83.26 [100, 60] loss: 0.207 [100, 120] loss: 0.225 [100, 180] loss: 0.211 [100, 240] loss: 0.217 [100, 300] loss: 0.219 [100, 360] loss: 0.223 Epoch: 100 -> Loss: 0.204636290669 Epoch: 100 -> Test Accuracy: 83.28 Finished Training [1, 60] loss: 1.644 [1, 120] loss: 0.821 [1, 180] loss: 0.744 [1, 240] loss: 0.710 [1, 300] loss: 0.679 [1, 360] loss: 0.649 Epoch: 1 -> Loss: 0.701272428036 Epoch: 1 -> Test Accuracy: 78.65 [2, 60] loss: 0.597 [2, 120] loss: 0.585 [2, 180] loss: 0.572 [2, 240] loss: 0.545 [2, 300] loss: 0.565 [2, 360] loss: 0.560 Epoch: 2 -> Loss: 0.561306715012 Epoch: 2 -> Test Accuracy: 80.37 [3, 60] loss: 0.513 [3, 120] loss: 0.520 [3, 180] loss: 0.511 [3, 240] loss: 0.508 [3, 300] loss: 0.510 [3, 360] loss: 0.508 Epoch: 3 -> Loss: 0.586759090424 Epoch: 3 -> Test Accuracy: 81.28 [4, 60] loss: 0.494 [4, 120] loss: 0.475 [4, 180] loss: 0.473 [4, 240] loss: 0.491 [4, 300] loss: 0.462 [4, 360] loss: 0.490 Epoch: 4 -> Loss: 0.434667438269 Epoch: 4 -> Test Accuracy: 81.58 [5, 60] loss: 0.442 [5, 120] loss: 0.460 [5, 180] loss: 0.451 [5, 240] loss: 0.461 [5, 300] loss: 0.440 [5, 360] loss: 0.465 Epoch: 5 -> Loss: 0.457592815161 Epoch: 5 -> Test Accuracy: 81.89 [6, 60] loss: 0.423 [6, 120] loss: 0.434 [6, 180] loss: 0.416 [6, 240] loss: 0.449 [6, 300] loss: 0.448 [6, 360] loss: 0.448 Epoch: 6 -> Loss: 0.407326281071 Epoch: 6 -> Test Accuracy: 83.07 [7, 60] loss: 0.428 [7, 120] loss: 0.404 [7, 180] loss: 0.411 [7, 240] loss: 0.436 [7, 300] loss: 0.428 [7, 360] loss: 0.439 Epoch: 7 -> Loss: 0.390617400408 Epoch: 7 -> Test Accuracy: 82.82 [8, 60] loss: 0.395 [8, 120] loss: 0.398 [8, 180] loss: 0.412 [8, 240] loss: 0.411 [8, 300] loss: 0.425 [8, 360] loss: 0.438 Epoch: 8 -> Loss: 0.387452274561 Epoch: 8 -> Test Accuracy: 83.57 [9, 60] loss: 0.402 [9, 120] loss: 0.409 [9, 180] loss: 0.389 [9, 240] loss: 0.402 [9, 300] loss: 0.417 [9, 360] loss: 0.426 Epoch: 9 -> Loss: 0.486477464437 Epoch: 9 -> Test Accuracy: 82.97 [10, 60] loss: 0.388 [10, 120] loss: 0.388 [10, 180] loss: 0.393 [10, 240] loss: 0.412 [10, 300] loss: 0.391 [10, 360] loss: 0.421 Epoch: 10 -> Loss: 0.426252067089 Epoch: 10 -> Test Accuracy: 84.33 [11, 60] loss: 0.380 [11, 120] loss: 0.385 [11, 180] loss: 0.388 [11, 240] loss: 0.404 [11, 300] loss: 0.397 [11, 360] loss: 0.404 Epoch: 11 -> Loss: 0.693639338017 Epoch: 11 -> Test Accuracy: 83.25 [12, 60] loss: 0.362 [12, 120] loss: 0.360 [12, 180] loss: 0.378 [12, 240] loss: 0.404 [12, 300] loss: 0.396 [12, 360] loss: 0.400 Epoch: 12 -> Loss: 0.53497749567 Epoch: 12 -> Test Accuracy: 83.23 [13, 60] loss: 0.369 [13, 120] loss: 0.372 [13, 180] loss: 0.364 [13, 240] loss: 0.381 [13, 300] loss: 0.404 [13, 360] loss: 0.414 Epoch: 13 -> Loss: 0.437541097403 Epoch: 13 -> Test Accuracy: 84.03 [14, 60] loss: 0.378 [14, 120] loss: 0.374 [14, 180] loss: 0.376 [14, 240] loss: 0.391 [14, 300] loss: 0.373 [14, 360] loss: 0.397 Epoch: 14 -> Loss: 0.38378995657 Epoch: 14 -> Test Accuracy: 83.58 [15, 60] loss: 0.370 [15, 120] loss: 0.372 [15, 180] loss: 0.377 [15, 240] loss: 0.386 [15, 300] loss: 0.364 [15, 360] loss: 0.381 Epoch: 15 -> Loss: 0.310917705297 Epoch: 15 -> Test Accuracy: 83.59 [16, 60] loss: 0.371 [16, 120] loss: 0.357 [16, 180] loss: 0.366 [16, 240] loss: 0.375 [16, 300] loss: 0.372 [16, 360] loss: 0.393 Epoch: 16 -> Loss: 0.276065915823 Epoch: 16 -> Test Accuracy: 83.46 [17, 60] loss: 0.357 [17, 120] loss: 0.356 [17, 180] loss: 0.381 [17, 240] loss: 0.373 [17, 300] loss: 0.382 [17, 360] loss: 0.379 Epoch: 17 -> Loss: 0.250696003437 Epoch: 17 -> Test Accuracy: 83.86 [18, 60] loss: 0.358 [18, 120] loss: 0.374 [18, 180] loss: 0.379 [18, 240] loss: 0.371 [18, 300] loss: 0.387 [18, 360] loss: 0.371 Epoch: 18 -> Loss: 0.264555454254 Epoch: 18 -> Test Accuracy: 83.73 [19, 60] loss: 0.344 [19, 120] loss: 0.359 [19, 180] loss: 0.371 [19, 240] loss: 0.373 [19, 300] loss: 0.379 [19, 360] loss: 0.382 Epoch: 19 -> Loss: 0.384335100651 Epoch: 19 -> Test Accuracy: 84.08 [20, 60] loss: 0.340 [20, 120] loss: 0.364 [20, 180] loss: 0.350 [20, 240] loss: 0.376 [20, 300] loss: 0.377 [20, 360] loss: 0.365 Epoch: 20 -> Loss: 0.480647653341 Epoch: 20 -> Test Accuracy: 83.36 [21, 60] loss: 0.303 [21, 120] loss: 0.312 [21, 180] loss: 0.289 [21, 240] loss: 0.288 [21, 300] loss: 0.295 [21, 360] loss: 0.273 Epoch: 21 -> Loss: 0.234928324819 Epoch: 21 -> Test Accuracy: 85.6 [22, 60] loss: 0.270 [22, 120] loss: 0.271 [22, 180] loss: 0.268 [22, 240] loss: 0.258 [22, 300] loss: 0.254 [22, 360] loss: 0.256 Epoch: 22 -> Loss: 0.315085470676 Epoch: 22 -> Test Accuracy: 85.78 [23, 60] loss: 0.243 [23, 120] loss: 0.246 [23, 180] loss: 0.241 [23, 240] loss: 0.244 [23, 300] loss: 0.248 [23, 360] loss: 0.240 Epoch: 23 -> Loss: 0.207869812846 Epoch: 23 -> Test Accuracy: 85.97 [24, 60] loss: 0.223 [24, 120] loss: 0.228 [24, 180] loss: 0.231 [24, 240] loss: 0.219 [24, 300] loss: 0.237 [24, 360] loss: 0.247 Epoch: 24 -> Loss: 0.22454893589 Epoch: 24 -> Test Accuracy: 86.11 [25, 60] loss: 0.215 [25, 120] loss: 0.224 [25, 180] loss: 0.227 [25, 240] loss: 0.216 [25, 300] loss: 0.225 [25, 360] loss: 0.228 Epoch: 25 -> Loss: 0.151886552572 Epoch: 25 -> Test Accuracy: 86.24 [26, 60] loss: 0.218 [26, 120] loss: 0.219 [26, 180] loss: 0.215 [26, 240] loss: 0.233 [26, 300] loss: 0.231 [26, 360] loss: 0.212 Epoch: 26 -> Loss: 0.146667078137 Epoch: 26 -> Test Accuracy: 86.05 [27, 60] loss: 0.197 [27, 120] loss: 0.212 [27, 180] loss: 0.211 [27, 240] loss: 0.213 [27, 300] loss: 0.216 [27, 360] loss: 0.221 Epoch: 27 -> Loss: 0.311142802238 Epoch: 27 -> Test Accuracy: 85.84 [28, 60] loss: 0.195 [28, 120] loss: 0.207 [28, 180] loss: 0.207 [28, 240] loss: 0.218 [28, 300] loss: 0.200 [28, 360] loss: 0.209 Epoch: 28 -> Loss: 0.142397254705 Epoch: 28 -> Test Accuracy: 85.58 [29, 60] loss: 0.192 [29, 120] loss: 0.209 [29, 180] loss: 0.199 [29, 240] loss: 0.206 [29, 300] loss: 0.199 [29, 360] loss: 0.204 Epoch: 29 -> Loss: 0.286255061626 Epoch: 29 -> Test Accuracy: 85.57 [30, 60] loss: 0.196 [30, 120] loss: 0.197 [30, 180] loss: 0.198 [30, 240] loss: 0.204 [30, 300] loss: 0.196 [30, 360] loss: 0.205 Epoch: 30 -> Loss: 0.151985600591 Epoch: 30 -> Test Accuracy: 85.61 [31, 60] loss: 0.185 [31, 120] loss: 0.205 [31, 180] loss: 0.200 [31, 240] loss: 0.197 [31, 300] loss: 0.200 [31, 360] loss: 0.219 Epoch: 31 -> Loss: 0.199593648314 Epoch: 31 -> Test Accuracy: 85.99 [32, 60] loss: 0.192 [32, 120] loss: 0.210 [32, 180] loss: 0.209 [32, 240] loss: 0.206 [32, 300] loss: 0.199 [32, 360] loss: 0.209 Epoch: 32 -> Loss: 0.191033646464 Epoch: 32 -> Test Accuracy: 85.83 [33, 60] loss: 0.194 [33, 120] loss: 0.191 [33, 180] loss: 0.190 [33, 240] loss: 0.189 [33, 300] loss: 0.195 [33, 360] loss: 0.201 Epoch: 33 -> Loss: 0.201527312398 Epoch: 33 -> Test Accuracy: 85.09 [34, 60] loss: 0.186 [34, 120] loss: 0.198 [34, 180] loss: 0.195 [34, 240] loss: 0.194 [34, 300] loss: 0.208 [34, 360] loss: 0.210 Epoch: 34 -> Loss: 0.233592748642 Epoch: 34 -> Test Accuracy: 85.53 [35, 60] loss: 0.181 [35, 120] loss: 0.179 [35, 180] loss: 0.187 [35, 240] loss: 0.201 [35, 300] loss: 0.206 [35, 360] loss: 0.201 Epoch: 35 -> Loss: 0.152111202478 Epoch: 35 -> Test Accuracy: 85.55 [36, 60] loss: 0.195 [36, 120] loss: 0.195 [36, 180] loss: 0.190 [36, 240] loss: 0.200 [36, 300] loss: 0.198 [36, 360] loss: 0.214 Epoch: 36 -> Loss: 0.131010040641 Epoch: 36 -> Test Accuracy: 85.71 [37, 60] loss: 0.189 [37, 120] loss: 0.182 [37, 180] loss: 0.182 [37, 240] loss: 0.190 [37, 300] loss: 0.198 [37, 360] loss: 0.203 Epoch: 37 -> Loss: 0.173107802868 Epoch: 37 -> Test Accuracy: 85.54 [38, 60] loss: 0.185 [38, 120] loss: 0.193 [38, 180] loss: 0.193 [38, 240] loss: 0.193 [38, 300] loss: 0.188 [38, 360] loss: 0.186 Epoch: 38 -> Loss: 0.176717355847 Epoch: 38 -> Test Accuracy: 85.57 [39, 60] loss: 0.180 [39, 120] loss: 0.182 [39, 180] loss: 0.195 [39, 240] loss: 0.185 [39, 300] loss: 0.198 [39, 360] loss: 0.210 Epoch: 39 -> Loss: 0.266058504581 Epoch: 39 -> Test Accuracy: 85.68 [40, 60] loss: 0.176 [40, 120] loss: 0.173 [40, 180] loss: 0.196 [40, 240] loss: 0.192 [40, 300] loss: 0.194 [40, 360] loss: 0.197 Epoch: 40 -> Loss: 0.314575463533 Epoch: 40 -> Test Accuracy: 85.45 [41, 60] loss: 0.163 [41, 120] loss: 0.159 [41, 180] loss: 0.164 [41, 240] loss: 0.152 [41, 300] loss: 0.157 [41, 360] loss: 0.145 Epoch: 41 -> Loss: 0.162493079901 Epoch: 41 -> Test Accuracy: 85.91 [42, 60] loss: 0.142 [42, 120] loss: 0.137 [42, 180] loss: 0.145 [42, 240] loss: 0.141 [42, 300] loss: 0.145 [42, 360] loss: 0.139 Epoch: 42 -> Loss: 0.162816256285 Epoch: 42 -> Test Accuracy: 86.22 [43, 60] loss: 0.125 [43, 120] loss: 0.131 [43, 180] loss: 0.131 [43, 240] loss: 0.138 [43, 300] loss: 0.141 [43, 360] loss: 0.142 Epoch: 43 -> Loss: 0.157318085432 Epoch: 43 -> Test Accuracy: 86.57 [44, 60] loss: 0.124 [44, 120] loss: 0.121 [44, 180] loss: 0.129 [44, 240] loss: 0.127 [44, 300] loss: 0.138 [44, 360] loss: 0.129 Epoch: 44 -> Loss: 0.0897534042597 Epoch: 44 -> Test Accuracy: 86.6 [45, 60] loss: 0.131 [45, 120] loss: 0.122 [45, 180] loss: 0.126 [45, 240] loss: 0.119 [45, 300] loss: 0.134 [45, 360] loss: 0.127 Epoch: 45 -> Loss: 0.0621313527226 Epoch: 45 -> Test Accuracy: 86.34 [46, 60] loss: 0.123 [46, 120] loss: 0.111 [46, 180] loss: 0.116 [46, 240] loss: 0.116 [46, 300] loss: 0.119 [46, 360] loss: 0.122 Epoch: 46 -> Loss: 0.121094107628 Epoch: 46 -> Test Accuracy: 86.45 [47, 60] loss: 0.116 [47, 120] loss: 0.119 [47, 180] loss: 0.114 [47, 240] loss: 0.115 [47, 300] loss: 0.116 [47, 360] loss: 0.115 Epoch: 47 -> Loss: 0.111577153206 Epoch: 47 -> Test Accuracy: 86.57 [48, 60] loss: 0.120 [48, 120] loss: 0.116 [48, 180] loss: 0.112 [48, 240] loss: 0.111 [48, 300] loss: 0.111 [48, 360] loss: 0.108 Epoch: 48 -> Loss: 0.214693546295 Epoch: 48 -> Test Accuracy: 86.52 [49, 60] loss: 0.105 [49, 120] loss: 0.112 [49, 180] loss: 0.111 [49, 240] loss: 0.111 [49, 300] loss: 0.114 [49, 360] loss: 0.110 Epoch: 49 -> Loss: 0.0762288421392 Epoch: 49 -> Test Accuracy: 86.5 [50, 60] loss: 0.110 [50, 120] loss: 0.109 [50, 180] loss: 0.112 [50, 240] loss: 0.114 [50, 300] loss: 0.107 [50, 360] loss: 0.112 Epoch: 50 -> Loss: 0.135717004538 Epoch: 50 -> Test Accuracy: 86.55 [51, 60] loss: 0.108 [51, 120] loss: 0.114 [51, 180] loss: 0.107 [51, 240] loss: 0.106 [51, 300] loss: 0.112 [51, 360] loss: 0.112 Epoch: 51 -> Loss: 0.0814069435 Epoch: 51 -> Test Accuracy: 86.54 [52, 60] loss: 0.111 [52, 120] loss: 0.111 [52, 180] loss: 0.107 [52, 240] loss: 0.108 [52, 300] loss: 0.108 [52, 360] loss: 0.114 Epoch: 52 -> Loss: 0.183041721582 Epoch: 52 -> Test Accuracy: 86.56 [53, 60] loss: 0.105 [53, 120] loss: 0.110 [53, 180] loss: 0.100 [53, 240] loss: 0.111 [53, 300] loss: 0.098 [53, 360] loss: 0.107 Epoch: 53 -> Loss: 0.111024282873 Epoch: 53 -> Test Accuracy: 86.5 [54, 60] loss: 0.102 [54, 120] loss: 0.109 [54, 180] loss: 0.099 [54, 240] loss: 0.106 [54, 300] loss: 0.105 [54, 360] loss: 0.109 Epoch: 54 -> Loss: 0.17025090754 Epoch: 54 -> Test Accuracy: 86.5 [55, 60] loss: 0.107 [55, 120] loss: 0.104 [55, 180] loss: 0.106 [55, 240] loss: 0.106 [55, 300] loss: 0.105 [55, 360] loss: 0.105 Epoch: 55 -> Loss: 0.123021617532 Epoch: 55 -> Test Accuracy: 86.49 [56, 60] loss: 0.099 [56, 120] loss: 0.101 [56, 180] loss: 0.105 [56, 240] loss: 0.100 [56, 300] loss: 0.105 [56, 360] loss: 0.102 Epoch: 56 -> Loss: 0.0429081134498 Epoch: 56 -> Test Accuracy: 86.66 [57, 60] loss: 0.099 [57, 120] loss: 0.100 [57, 180] loss: 0.097 [57, 240] loss: 0.102 [57, 300] loss: 0.099 [57, 360] loss: 0.101 Epoch: 57 -> Loss: 0.085507825017 Epoch: 57 -> Test Accuracy: 86.43 [58, 60] loss: 0.105 [58, 120] loss: 0.102 [58, 180] loss: 0.107 [58, 240] loss: 0.095 [58, 300] loss: 0.094 [58, 360] loss: 0.101 Epoch: 58 -> Loss: 0.186266377568 Epoch: 58 -> Test Accuracy: 86.62 [59, 60] loss: 0.101 [59, 120] loss: 0.105 [59, 180] loss: 0.098 [59, 240] loss: 0.099 [59, 300] loss: 0.102 [59, 360] loss: 0.102 Epoch: 59 -> Loss: 0.116454288363 Epoch: 59 -> Test Accuracy: 86.42 [60, 60] loss: 0.110 [60, 120] loss: 0.096 [60, 180] loss: 0.101 [60, 240] loss: 0.101 [60, 300] loss: 0.101 [60, 360] loss: 0.101 Epoch: 60 -> Loss: 0.199208289385 Epoch: 60 -> Test Accuracy: 86.56 [61, 60] loss: 0.099 [61, 120] loss: 0.089 [61, 180] loss: 0.101 [61, 240] loss: 0.097 [61, 300] loss: 0.100 [61, 360] loss: 0.096 Epoch: 61 -> Loss: 0.229684591293 Epoch: 61 -> Test Accuracy: 86.53 [62, 60] loss: 0.093 [62, 120] loss: 0.100 [62, 180] loss: 0.096 [62, 240] loss: 0.091 [62, 300] loss: 0.105 [62, 360] loss: 0.101 Epoch: 62 -> Loss: 0.0992732420564 Epoch: 62 -> Test Accuracy: 86.34 [63, 60] loss: 0.098 [63, 120] loss: 0.092 [63, 180] loss: 0.096 [63, 240] loss: 0.095 [63, 300] loss: 0.097 [63, 360] loss: 0.093 Epoch: 63 -> Loss: 0.107236407697 Epoch: 63 -> Test Accuracy: 86.44 [64, 60] loss: 0.093 [64, 120] loss: 0.096 [64, 180] loss: 0.097 [64, 240] loss: 0.096 [64, 300] loss: 0.095 [64, 360] loss: 0.099 Epoch: 64 -> Loss: 0.141364723444 Epoch: 64 -> Test Accuracy: 86.53 [65, 60] loss: 0.097 [65, 120] loss: 0.094 [65, 180] loss: 0.103 [65, 240] loss: 0.086 [65, 300] loss: 0.089 [65, 360] loss: 0.107 Epoch: 65 -> Loss: 0.11465189606 Epoch: 65 -> Test Accuracy: 86.36 [66, 60] loss: 0.099 [66, 120] loss: 0.098 [66, 180] loss: 0.092 [66, 240] loss: 0.088 [66, 300] loss: 0.097 [66, 360] loss: 0.091 Epoch: 66 -> Loss: 0.126636952162 Epoch: 66 -> Test Accuracy: 86.49 [67, 60] loss: 0.090 [67, 120] loss: 0.089 [67, 180] loss: 0.094 [67, 240] loss: 0.091 [67, 300] loss: 0.091 [67, 360] loss: 0.097 Epoch: 67 -> Loss: 0.0722390115261 Epoch: 67 -> Test Accuracy: 86.52 [68, 60] loss: 0.093 [68, 120] loss: 0.090 [68, 180] loss: 0.093 [68, 240] loss: 0.095 [68, 300] loss: 0.094 [68, 360] loss: 0.097 Epoch: 68 -> Loss: 0.0976048186421 Epoch: 68 -> Test Accuracy: 86.5 [69, 60] loss: 0.090 [69, 120] loss: 0.094 [69, 180] loss: 0.090 [69, 240] loss: 0.098 [69, 300] loss: 0.087 [69, 360] loss: 0.090 Epoch: 69 -> Loss: 0.124031342566 Epoch: 69 -> Test Accuracy: 86.45 [70, 60] loss: 0.094 [70, 120] loss: 0.088 [70, 180] loss: 0.090 [70, 240] loss: 0.087 [70, 300] loss: 0.089 [70, 360] loss: 0.091 Epoch: 70 -> Loss: 0.0749558657408 Epoch: 70 -> Test Accuracy: 86.42 [71, 60] loss: 0.092 [71, 120] loss: 0.093 [71, 180] loss: 0.086 [71, 240] loss: 0.088 [71, 300] loss: 0.084 [71, 360] loss: 0.094 Epoch: 71 -> Loss: 0.264926105738 Epoch: 71 -> Test Accuracy: 86.61 [72, 60] loss: 0.092 [72, 120] loss: 0.083 [72, 180] loss: 0.090 [72, 240] loss: 0.089 [72, 300] loss: 0.087 [72, 360] loss: 0.090 Epoch: 72 -> Loss: 0.0532714352012 Epoch: 72 -> Test Accuracy: 86.57 [73, 60] loss: 0.094 [73, 120] loss: 0.092 [73, 180] loss: 0.088 [73, 240] loss: 0.092 [73, 300] loss: 0.083 [73, 360] loss: 0.090 Epoch: 73 -> Loss: 0.0466169789433 Epoch: 73 -> Test Accuracy: 86.54 [74, 60] loss: 0.090 [74, 120] loss: 0.087 [74, 180] loss: 0.086 [74, 240] loss: 0.092 [74, 300] loss: 0.086 [74, 360] loss: 0.088 Epoch: 74 -> Loss: 0.0855236947536 Epoch: 74 -> Test Accuracy: 86.55 [75, 60] loss: 0.086 [75, 120] loss: 0.088 [75, 180] loss: 0.084 [75, 240] loss: 0.090 [75, 300] loss: 0.084 [75, 360] loss: 0.087 Epoch: 75 -> Loss: 0.0465160682797 Epoch: 75 -> Test Accuracy: 86.55 [76, 60] loss: 0.094 [76, 120] loss: 0.088 [76, 180] loss: 0.081 [76, 240] loss: 0.087 [76, 300] loss: 0.090 [76, 360] loss: 0.093 Epoch: 76 -> Loss: 0.0497290790081 Epoch: 76 -> Test Accuracy: 86.56 [77, 60] loss: 0.084 [77, 120] loss: 0.089 [77, 180] loss: 0.088 [77, 240] loss: 0.080 [77, 300] loss: 0.088 [77, 360] loss: 0.088 Epoch: 77 -> Loss: 0.0906545221806 Epoch: 77 -> Test Accuracy: 86.48 [78, 60] loss: 0.084 [78, 120] loss: 0.084 [78, 180] loss: 0.085 [78, 240] loss: 0.083 [78, 300] loss: 0.085 [78, 360] loss: 0.094 Epoch: 78 -> Loss: 0.121591173112 Epoch: 78 -> Test Accuracy: 86.69 [79, 60] loss: 0.078 [79, 120] loss: 0.084 [79, 180] loss: 0.083 [79, 240] loss: 0.087 [79, 300] loss: 0.094 [79, 360] loss: 0.086 Epoch: 79 -> Loss: 0.142673268914 Epoch: 79 -> Test Accuracy: 86.56 [80, 60] loss: 0.088 [80, 120] loss: 0.083 [80, 180] loss: 0.086 [80, 240] loss: 0.090 [80, 300] loss: 0.079 [80, 360] loss: 0.082 Epoch: 80 -> Loss: 0.0883778780699 Epoch: 80 -> Test Accuracy: 86.63 [81, 60] loss: 0.084 [81, 120] loss: 0.085 [81, 180] loss: 0.081 [81, 240] loss: 0.085 [81, 300] loss: 0.083 [81, 360] loss: 0.087 Epoch: 81 -> Loss: 0.0469513982534 Epoch: 81 -> Test Accuracy: 86.6 [82, 60] loss: 0.088 [82, 120] loss: 0.083 [82, 180] loss: 0.082 [82, 240] loss: 0.088 [82, 300] loss: 0.083 [82, 360] loss: 0.087 Epoch: 82 -> Loss: 0.0524021983147 Epoch: 82 -> Test Accuracy: 86.54 [83, 60] loss: 0.086 [83, 120] loss: 0.086 [83, 180] loss: 0.084 [83, 240] loss: 0.079 [83, 300] loss: 0.078 [83, 360] loss: 0.088 Epoch: 83 -> Loss: 0.116114482284 Epoch: 83 -> Test Accuracy: 86.54 [84, 60] loss: 0.075 [84, 120] loss: 0.085 [84, 180] loss: 0.078 [84, 240] loss: 0.078 [84, 300] loss: 0.085 [84, 360] loss: 0.081 Epoch: 84 -> Loss: 0.0494220852852 Epoch: 84 -> Test Accuracy: 86.39 [85, 60] loss: 0.085 [85, 120] loss: 0.080 [85, 180] loss: 0.085 [85, 240] loss: 0.081 [85, 300] loss: 0.085 [85, 360] loss: 0.085 Epoch: 85 -> Loss: 0.0958910509944 Epoch: 85 -> Test Accuracy: 86.6 [86, 60] loss: 0.074 [86, 120] loss: 0.083 [86, 180] loss: 0.081 [86, 240] loss: 0.083 [86, 300] loss: 0.080 [86, 360] loss: 0.086 Epoch: 86 -> Loss: 0.0433759093285 Epoch: 86 -> Test Accuracy: 86.43 [87, 60] loss: 0.079 [87, 120] loss: 0.084 [87, 180] loss: 0.083 [87, 240] loss: 0.088 [87, 300] loss: 0.080 [87, 360] loss: 0.080 Epoch: 87 -> Loss: 0.0512029603124 Epoch: 87 -> Test Accuracy: 86.61 [88, 60] loss: 0.077 [88, 120] loss: 0.082 [88, 180] loss: 0.082 [88, 240] loss: 0.079 [88, 300] loss: 0.079 [88, 360] loss: 0.081 Epoch: 88 -> Loss: 0.142117246985 Epoch: 88 -> Test Accuracy: 86.45 [89, 60] loss: 0.082 [89, 120] loss: 0.078 [89, 180] loss: 0.077 [89, 240] loss: 0.078 [89, 300] loss: 0.080 [89, 360] loss: 0.078 Epoch: 89 -> Loss: 0.0750356838107 Epoch: 89 -> Test Accuracy: 86.61 [90, 60] loss: 0.078 [90, 120] loss: 0.077 [90, 180] loss: 0.079 [90, 240] loss: 0.079 [90, 300] loss: 0.076 [90, 360] loss: 0.075 Epoch: 90 -> Loss: 0.0498711057007 Epoch: 90 -> Test Accuracy: 86.61 [91, 60] loss: 0.080 [91, 120] loss: 0.077 [91, 180] loss: 0.077 [91, 240] loss: 0.077 [91, 300] loss: 0.075 [91, 360] loss: 0.080 Epoch: 91 -> Loss: 0.189132228494 Epoch: 91 -> Test Accuracy: 86.5 [92, 60] loss: 0.075 [92, 120] loss: 0.074 [92, 180] loss: 0.075 [92, 240] loss: 0.076 [92, 300] loss: 0.085 [92, 360] loss: 0.076 Epoch: 92 -> Loss: 0.0656943097711 Epoch: 92 -> Test Accuracy: 86.42 [93, 60] loss: 0.075 [93, 120] loss: 0.071 [93, 180] loss: 0.076 [93, 240] loss: 0.076 [93, 300] loss: 0.076 [93, 360] loss: 0.075 Epoch: 93 -> Loss: 0.106649264693 Epoch: 93 -> Test Accuracy: 86.51 [94, 60] loss: 0.088 [94, 120] loss: 0.072 [94, 180] loss: 0.072 [94, 240] loss: 0.074 [94, 300] loss: 0.071 [94, 360] loss: 0.081 Epoch: 94 -> Loss: 0.119260944426 Epoch: 94 -> Test Accuracy: 86.52 [95, 60] loss: 0.076 [95, 120] loss: 0.074 [95, 180] loss: 0.078 [95, 240] loss: 0.073 [95, 300] loss: 0.076 [95, 360] loss: 0.076 Epoch: 95 -> Loss: 0.0358727164567 Epoch: 95 -> Test Accuracy: 86.48 [96, 60] loss: 0.077 [96, 120] loss: 0.069 [96, 180] loss: 0.074 [96, 240] loss: 0.079 [96, 300] loss: 0.075 [96, 360] loss: 0.071 Epoch: 96 -> Loss: 0.103197000921 Epoch: 96 -> Test Accuracy: 86.7 [97, 60] loss: 0.074 [97, 120] loss: 0.075 [97, 180] loss: 0.071 [97, 240] loss: 0.079 [97, 300] loss: 0.073 [97, 360] loss: 0.076 Epoch: 97 -> Loss: 0.157634466887 Epoch: 97 -> Test Accuracy: 86.58 [98, 60] loss: 0.078 [98, 120] loss: 0.073 [98, 180] loss: 0.066 [98, 240] loss: 0.077 [98, 300] loss: 0.078 [98, 360] loss: 0.071 Epoch: 98 -> Loss: 0.107171714306 Epoch: 98 -> Test Accuracy: 86.53 [99, 60] loss: 0.069 [99, 120] loss: 0.075 [99, 180] loss: 0.081 [99, 240] loss: 0.074 [99, 300] loss: 0.077 [99, 360] loss: 0.068 Epoch: 99 -> Loss: 0.254891574383 Epoch: 99 -> Test Accuracy: 86.6 [100, 60] loss: 0.074 [100, 120] loss: 0.073 [100, 180] loss: 0.076 [100, 240] loss: 0.077 [100, 300] loss: 0.078 [100, 360] loss: 0.078 Epoch: 100 -> Loss: 0.0800180584192 Epoch: 100 -> Test Accuracy: 86.51 Finished Training [1, 60] loss: 2.728 [1, 120] loss: 1.803 [1, 180] loss: 1.762 [1, 240] loss: 1.721 [1, 300] loss: 1.705 [1, 360] loss: 1.687 Epoch: 1 -> Loss: 1.52180790901 Epoch: 1 -> Test Accuracy: 37.26 [2, 60] loss: 1.664 [2, 120] loss: 1.644 [2, 180] loss: 1.637 [2, 240] loss: 1.625 [2, 300] loss: 1.591 [2, 360] loss: 1.606 Epoch: 2 -> Loss: 1.53919839859 Epoch: 2 -> Test Accuracy: 40.21 [3, 60] loss: 1.565 [3, 120] loss: 1.564 [3, 180] loss: 1.590 [3, 240] loss: 1.567 [3, 300] loss: 1.561 [3, 360] loss: 1.553 Epoch: 3 -> Loss: 1.50751459599 Epoch: 3 -> Test Accuracy: 40.98 [4, 60] loss: 1.559 [4, 120] loss: 1.531 [4, 180] loss: 1.526 [4, 240] loss: 1.527 [4, 300] loss: 1.523 [4, 360] loss: 1.516 Epoch: 4 -> Loss: 1.84346234798 Epoch: 4 -> Test Accuracy: 42.62 [5, 60] loss: 1.516 [5, 120] loss: 1.522 [5, 180] loss: 1.520 [5, 240] loss: 1.507 [5, 300] loss: 1.493 [5, 360] loss: 1.504 Epoch: 5 -> Loss: 1.55798828602 Epoch: 5 -> Test Accuracy: 43.2 [6, 60] loss: 1.506 [6, 120] loss: 1.501 [6, 180] loss: 1.493 [6, 240] loss: 1.502 [6, 300] loss: 1.496 [6, 360] loss: 1.489 Epoch: 6 -> Loss: 1.55128777027 Epoch: 6 -> Test Accuracy: 44.04 [7, 60] loss: 1.484 [7, 120] loss: 1.479 [7, 180] loss: 1.480 [7, 240] loss: 1.489 [7, 300] loss: 1.490 [7, 360] loss: 1.497 Epoch: 7 -> Loss: 1.38384413719 Epoch: 7 -> Test Accuracy: 43.72 [8, 60] loss: 1.477 [8, 120] loss: 1.487 [8, 180] loss: 1.476 [8, 240] loss: 1.464 [8, 300] loss: 1.475 [8, 360] loss: 1.473 Epoch: 8 -> Loss: 1.44293642044 Epoch: 8 -> Test Accuracy: 43.53 [9, 60] loss: 1.465 [9, 120] loss: 1.465 [9, 180] loss: 1.478 [9, 240] loss: 1.463 [9, 300] loss: 1.478 [9, 360] loss: 1.462 Epoch: 9 -> Loss: 1.58194768429 Epoch: 9 -> Test Accuracy: 44.31 [10, 60] loss: 1.480 [10, 120] loss: 1.461 [10, 180] loss: 1.447 [10, 240] loss: 1.470 [10, 300] loss: 1.470 [10, 360] loss: 1.457 Epoch: 10 -> Loss: 1.58967161179 Epoch: 10 -> Test Accuracy: 44.39 [11, 60] loss: 1.453 [11, 120] loss: 1.458 [11, 180] loss: 1.471 [11, 240] loss: 1.458 [11, 300] loss: 1.453 [11, 360] loss: 1.481 Epoch: 11 -> Loss: 1.41367149353 Epoch: 11 -> Test Accuracy: 43.93 [12, 60] loss: 1.469 [12, 120] loss: 1.442 [12, 180] loss: 1.453 [12, 240] loss: 1.462 [12, 300] loss: 1.472 [12, 360] loss: 1.448 Epoch: 12 -> Loss: 1.50912034512 Epoch: 12 -> Test Accuracy: 45.51 [13, 60] loss: 1.460 [13, 120] loss: 1.454 [13, 180] loss: 1.432 [13, 240] loss: 1.463 [13, 300] loss: 1.443 [13, 360] loss: 1.472 Epoch: 13 -> Loss: 1.46690046787 Epoch: 13 -> Test Accuracy: 44.22 [14, 60] loss: 1.442 [14, 120] loss: 1.456 [14, 180] loss: 1.473 [14, 240] loss: 1.452 [14, 300] loss: 1.442 [14, 360] loss: 1.444 Epoch: 14 -> Loss: 1.42689204216 Epoch: 14 -> Test Accuracy: 44.5 [15, 60] loss: 1.440 [15, 120] loss: 1.459 [15, 180] loss: 1.451 [15, 240] loss: 1.440 [15, 300] loss: 1.465 [15, 360] loss: 1.448 Epoch: 15 -> Loss: 1.5370644331 Epoch: 15 -> Test Accuracy: 45.04 [16, 60] loss: 1.444 [16, 120] loss: 1.434 [16, 180] loss: 1.449 [16, 240] loss: 1.461 [16, 300] loss: 1.434 [16, 360] loss: 1.467 Epoch: 16 -> Loss: 1.54545843601 Epoch: 16 -> Test Accuracy: 45.42 [17, 60] loss: 1.467 [17, 120] loss: 1.426 [17, 180] loss: 1.440 [17, 240] loss: 1.458 [17, 300] loss: 1.439 [17, 360] loss: 1.445 Epoch: 17 -> Loss: 1.40365207195 Epoch: 17 -> Test Accuracy: 45.07 [18, 60] loss: 1.464 [18, 120] loss: 1.444 [18, 180] loss: 1.432 [18, 240] loss: 1.442 [18, 300] loss: 1.439 [18, 360] loss: 1.430 Epoch: 18 -> Loss: 1.54345631599 Epoch: 18 -> Test Accuracy: 45.02 [19, 60] loss: 1.436 [19, 120] loss: 1.438 [19, 180] loss: 1.449 [19, 240] loss: 1.448 [19, 300] loss: 1.452 [19, 360] loss: 1.458 Epoch: 19 -> Loss: 1.54569327831 Epoch: 19 -> Test Accuracy: 44.76 [20, 60] loss: 1.443 [20, 120] loss: 1.439 [20, 180] loss: 1.439 [20, 240] loss: 1.440 [20, 300] loss: 1.434 [20, 360] loss: 1.445 Epoch: 20 -> Loss: 1.4499450922 Epoch: 20 -> Test Accuracy: 44.51 [21, 60] loss: 1.390 [21, 120] loss: 1.365 [21, 180] loss: 1.379 [21, 240] loss: 1.349 [21, 300] loss: 1.343 [21, 360] loss: 1.330 Epoch: 21 -> Loss: 1.20579361916 Epoch: 21 -> Test Accuracy: 48.52 [22, 60] loss: 1.311 [22, 120] loss: 1.325 [22, 180] loss: 1.327 [22, 240] loss: 1.327 [22, 300] loss: 1.317 [22, 360] loss: 1.318 Epoch: 22 -> Loss: 1.28634214401 Epoch: 22 -> Test Accuracy: 49.25 [23, 60] loss: 1.311 [23, 120] loss: 1.300 [23, 180] loss: 1.323 [23, 240] loss: 1.301 [23, 300] loss: 1.323 [23, 360] loss: 1.313 Epoch: 23 -> Loss: 1.2557246685 Epoch: 23 -> Test Accuracy: 48.96 [24, 60] loss: 1.288 [24, 120] loss: 1.296 [24, 180] loss: 1.312 [24, 240] loss: 1.288 [24, 300] loss: 1.303 [24, 360] loss: 1.276 Epoch: 24 -> Loss: 1.45214509964 Epoch: 24 -> Test Accuracy: 49.58 [25, 60] loss: 1.285 [25, 120] loss: 1.310 [25, 180] loss: 1.301 [25, 240] loss: 1.305 [25, 300] loss: 1.289 [25, 360] loss: 1.311 Epoch: 25 -> Loss: 1.18165540695 Epoch: 25 -> Test Accuracy: 49.85 [26, 60] loss: 1.289 [26, 120] loss: 1.295 [26, 180] loss: 1.272 [26, 240] loss: 1.297 [26, 300] loss: 1.296 [26, 360] loss: 1.305 Epoch: 26 -> Loss: 1.342638731 Epoch: 26 -> Test Accuracy: 49.52 [27, 60] loss: 1.299 [27, 120] loss: 1.287 [27, 180] loss: 1.279 [27, 240] loss: 1.294 [27, 300] loss: 1.299 [27, 360] loss: 1.288 Epoch: 27 -> Loss: 1.19322669506 Epoch: 27 -> Test Accuracy: 49.5 [28, 60] loss: 1.267 [28, 120] loss: 1.301 [28, 180] loss: 1.282 [28, 240] loss: 1.282 [28, 300] loss: 1.300 [28, 360] loss: 1.298 Epoch: 28 -> Loss: 1.40955781937 Epoch: 28 -> Test Accuracy: 48.96 [29, 60] loss: 1.263 [29, 120] loss: 1.296 [29, 180] loss: 1.287 [29, 240] loss: 1.306 [29, 300] loss: 1.296 [29, 360] loss: 1.284 Epoch: 29 -> Loss: 1.24729180336 Epoch: 29 -> Test Accuracy: 49.11 [30, 60] loss: 1.292 [30, 120] loss: 1.310 [30, 180] loss: 1.291 [30, 240] loss: 1.283 [30, 300] loss: 1.289 [30, 360] loss: 1.281 Epoch: 30 -> Loss: 1.07997095585 Epoch: 30 -> Test Accuracy: 49.38 [31, 60] loss: 1.292 [31, 120] loss: 1.283 [31, 180] loss: 1.282 [31, 240] loss: 1.289 [31, 300] loss: 1.289 [31, 360] loss: 1.290 Epoch: 31 -> Loss: 1.2045545578 Epoch: 31 -> Test Accuracy: 50.4 [32, 60] loss: 1.290 [32, 120] loss: 1.273 [32, 180] loss: 1.266 [32, 240] loss: 1.285 [32, 300] loss: 1.307 [32, 360] loss: 1.281 Epoch: 32 -> Loss: 1.08216369152 Epoch: 32 -> Test Accuracy: 49.56 [33, 60] loss: 1.251 [33, 120] loss: 1.296 [33, 180] loss: 1.269 [33, 240] loss: 1.302 [33, 300] loss: 1.294 [33, 360] loss: 1.298 Epoch: 33 -> Loss: 1.24276268482 Epoch: 33 -> Test Accuracy: 49.55 [34, 60] loss: 1.265 [34, 120] loss: 1.276 [34, 180] loss: 1.285 [34, 240] loss: 1.277 [34, 300] loss: 1.275 [34, 360] loss: 1.284 Epoch: 34 -> Loss: 1.29300177097 Epoch: 34 -> Test Accuracy: 50.31 [35, 60] loss: 1.271 [35, 120] loss: 1.293 [35, 180] loss: 1.297 [35, 240] loss: 1.288 [35, 300] loss: 1.273 [35, 360] loss: 1.282 Epoch: 35 -> Loss: 1.3820245266 Epoch: 35 -> Test Accuracy: 49.5 [36, 60] loss: 1.278 [36, 120] loss: 1.281 [36, 180] loss: 1.269 [36, 240] loss: 1.286 [36, 300] loss: 1.298 [36, 360] loss: 1.284 Epoch: 36 -> Loss: 1.44006991386 Epoch: 36 -> Test Accuracy: 49.84 [37, 60] loss: 1.275 [37, 120] loss: 1.285 [37, 180] loss: 1.307 [37, 240] loss: 1.279 [37, 300] loss: 1.267 [37, 360] loss: 1.281 Epoch: 37 -> Loss: 1.30595421791 Epoch: 37 -> Test Accuracy: 49.44 [38, 60] loss: 1.291 [38, 120] loss: 1.258 [38, 180] loss: 1.266 [38, 240] loss: 1.292 [38, 300] loss: 1.259 [38, 360] loss: 1.290 Epoch: 38 -> Loss: 1.15272068977 Epoch: 38 -> Test Accuracy: 48.94 [39, 60] loss: 1.287 [39, 120] loss: 1.297 [39, 180] loss: 1.273 [39, 240] loss: 1.272 [39, 300] loss: 1.260 [39, 360] loss: 1.288 Epoch: 39 -> Loss: 1.43694400787 Epoch: 39 -> Test Accuracy: 50.06 [40, 60] loss: 1.276 [40, 120] loss: 1.264 [40, 180] loss: 1.261 [40, 240] loss: 1.286 [40, 300] loss: 1.279 [40, 360] loss: 1.279 Epoch: 40 -> Loss: 1.37329339981 Epoch: 40 -> Test Accuracy: 49.91 [41, 60] loss: 1.243 [41, 120] loss: 1.235 [41, 180] loss: 1.220 [41, 240] loss: 1.211 [41, 300] loss: 1.230 [41, 360] loss: 1.229 Epoch: 41 -> Loss: 1.22143793106 Epoch: 41 -> Test Accuracy: 52.09 [42, 60] loss: 1.219 [42, 120] loss: 1.216 [42, 180] loss: 1.210 [42, 240] loss: 1.188 [42, 300] loss: 1.198 [42, 360] loss: 1.196 Epoch: 42 -> Loss: 1.20635735989 Epoch: 42 -> Test Accuracy: 52.29 [43, 60] loss: 1.213 [43, 120] loss: 1.209 [43, 180] loss: 1.196 [43, 240] loss: 1.181 [43, 300] loss: 1.180 [43, 360] loss: 1.197 Epoch: 43 -> Loss: 1.05017518997 Epoch: 43 -> Test Accuracy: 52.48 [44, 60] loss: 1.185 [44, 120] loss: 1.188 [44, 180] loss: 1.186 [44, 240] loss: 1.203 [44, 300] loss: 1.192 [44, 360] loss: 1.204 Epoch: 44 -> Loss: 1.16620528698 Epoch: 44 -> Test Accuracy: 52.36 [45, 60] loss: 1.197 [45, 120] loss: 1.202 [45, 180] loss: 1.183 [45, 240] loss: 1.192 [45, 300] loss: 1.181 [45, 360] loss: 1.180 Epoch: 45 -> Loss: 1.24895977974 Epoch: 45 -> Test Accuracy: 52.13 [46, 60] loss: 1.174 [46, 120] loss: 1.196 [46, 180] loss: 1.174 [46, 240] loss: 1.178 [46, 300] loss: 1.176 [46, 360] loss: 1.187 Epoch: 46 -> Loss: 1.15635371208 Epoch: 46 -> Test Accuracy: 52.73 [47, 60] loss: 1.169 [47, 120] loss: 1.173 [47, 180] loss: 1.158 [47, 240] loss: 1.159 [47, 300] loss: 1.186 [47, 360] loss: 1.170 Epoch: 47 -> Loss: 1.25530695915 Epoch: 47 -> Test Accuracy: 52.82 [48, 60] loss: 1.161 [48, 120] loss: 1.172 [48, 180] loss: 1.148 [48, 240] loss: 1.169 [48, 300] loss: 1.155 [48, 360] loss: 1.173 Epoch: 48 -> Loss: 1.15647470951 Epoch: 48 -> Test Accuracy: 53.03 [49, 60] loss: 1.180 [49, 120] loss: 1.167 [49, 180] loss: 1.144 [49, 240] loss: 1.184 [49, 300] loss: 1.153 [49, 360] loss: 1.154 Epoch: 49 -> Loss: 1.02462029457 Epoch: 49 -> Test Accuracy: 53.06 [50, 60] loss: 1.152 [50, 120] loss: 1.154 [50, 180] loss: 1.148 [50, 240] loss: 1.178 [50, 300] loss: 1.159 [50, 360] loss: 1.159 Epoch: 50 -> Loss: 1.01887559891 Epoch: 50 -> Test Accuracy: 53.1 [51, 60] loss: 1.158 [51, 120] loss: 1.145 [51, 180] loss: 1.157 [51, 240] loss: 1.177 [51, 300] loss: 1.156 [51, 360] loss: 1.163 Epoch: 51 -> Loss: 1.36545753479 Epoch: 51 -> Test Accuracy: 53.01 [52, 60] loss: 1.170 [52, 120] loss: 1.165 [52, 180] loss: 1.155 [52, 240] loss: 1.168 [52, 300] loss: 1.157 [52, 360] loss: 1.160 Epoch: 52 -> Loss: 1.01275980473 Epoch: 52 -> Test Accuracy: 53.23 [53, 60] loss: 1.151 [53, 120] loss: 1.154 [53, 180] loss: 1.165 [53, 240] loss: 1.142 [53, 300] loss: 1.161 [53, 360] loss: 1.172 Epoch: 53 -> Loss: 1.24386143684 Epoch: 53 -> Test Accuracy: 53.2 [54, 60] loss: 1.126 [54, 120] loss: 1.158 [54, 180] loss: 1.130 [54, 240] loss: 1.156 [54, 300] loss: 1.164 [54, 360] loss: 1.165 Epoch: 54 -> Loss: 1.104996562 Epoch: 54 -> Test Accuracy: 53.26 [55, 60] loss: 1.175 [55, 120] loss: 1.154 [55, 180] loss: 1.152 [55, 240] loss: 1.156 [55, 300] loss: 1.136 [55, 360] loss: 1.138 Epoch: 55 -> Loss: 1.11907505989 Epoch: 55 -> Test Accuracy: 53.21 [56, 60] loss: 1.147 [56, 120] loss: 1.183 [56, 180] loss: 1.155 [56, 240] loss: 1.149 [56, 300] loss: 1.156 [56, 360] loss: 1.154 Epoch: 56 -> Loss: 1.12735331059 Epoch: 56 -> Test Accuracy: 53.32 [57, 60] loss: 1.143 [57, 120] loss: 1.138 [57, 180] loss: 1.153 [57, 240] loss: 1.162 [57, 300] loss: 1.154 [57, 360] loss: 1.143 Epoch: 57 -> Loss: 1.07067346573 Epoch: 57 -> Test Accuracy: 53.12 [58, 60] loss: 1.150 [58, 120] loss: 1.152 [58, 180] loss: 1.143 [58, 240] loss: 1.157 [58, 300] loss: 1.154 [58, 360] loss: 1.156 Epoch: 58 -> Loss: 1.20118761063 Epoch: 58 -> Test Accuracy: 53.73 [59, 60] loss: 1.155 [59, 120] loss: 1.143 [59, 180] loss: 1.137 [59, 240] loss: 1.162 [59, 300] loss: 1.144 [59, 360] loss: 1.154 Epoch: 59 -> Loss: 1.18427968025 Epoch: 59 -> Test Accuracy: 53.41 [60, 60] loss: 1.153 [60, 120] loss: 1.148 [60, 180] loss: 1.152 [60, 240] loss: 1.149 [60, 300] loss: 1.133 [60, 360] loss: 1.184 Epoch: 60 -> Loss: 1.13069331646 Epoch: 60 -> Test Accuracy: 53.65 [61, 60] loss: 1.161 [61, 120] loss: 1.144 [61, 180] loss: 1.146 [61, 240] loss: 1.138 [61, 300] loss: 1.159 [61, 360] loss: 1.142 Epoch: 61 -> Loss: 1.14365255833 Epoch: 61 -> Test Accuracy: 53.59 [62, 60] loss: 1.143 [62, 120] loss: 1.140 [62, 180] loss: 1.155 [62, 240] loss: 1.138 [62, 300] loss: 1.152 [62, 360] loss: 1.145 Epoch: 62 -> Loss: 1.13571286201 Epoch: 62 -> Test Accuracy: 53.6 [63, 60] loss: 1.148 [63, 120] loss: 1.157 [63, 180] loss: 1.155 [63, 240] loss: 1.143 [63, 300] loss: 1.128 [63, 360] loss: 1.150 Epoch: 63 -> Loss: 1.21636080742 Epoch: 63 -> Test Accuracy: 53.39 [64, 60] loss: 1.145 [64, 120] loss: 1.160 [64, 180] loss: 1.127 [64, 240] loss: 1.149 [64, 300] loss: 1.145 [64, 360] loss: 1.168 Epoch: 64 -> Loss: 1.04232823849 Epoch: 64 -> Test Accuracy: 53.85 [65, 60] loss: 1.141 [65, 120] loss: 1.139 [65, 180] loss: 1.148 [65, 240] loss: 1.135 [65, 300] loss: 1.136 [65, 360] loss: 1.153 Epoch: 65 -> Loss: 1.18509340286 Epoch: 65 -> Test Accuracy: 53.51 [66, 60] loss: 1.137 [66, 120] loss: 1.151 [66, 180] loss: 1.158 [66, 240] loss: 1.136 [66, 300] loss: 1.131 [66, 360] loss: 1.143 Epoch: 66 -> Loss: 1.11153292656 Epoch: 66 -> Test Accuracy: 53.58 [67, 60] loss: 1.126 [67, 120] loss: 1.166 [67, 180] loss: 1.139 [67, 240] loss: 1.155 [67, 300] loss: 1.129 [67, 360] loss: 1.154 Epoch: 67 -> Loss: 1.34183907509 Epoch: 67 -> Test Accuracy: 53.62 [68, 60] loss: 1.153 [68, 120] loss: 1.138 [68, 180] loss: 1.126 [68, 240] loss: 1.159 [68, 300] loss: 1.139 [68, 360] loss: 1.133 Epoch: 68 -> Loss: 1.1199092865 Epoch: 68 -> Test Accuracy: 53.69 [69, 60] loss: 1.148 [69, 120] loss: 1.126 [69, 180] loss: 1.155 [69, 240] loss: 1.129 [69, 300] loss: 1.150 [69, 360] loss: 1.146 Epoch: 69 -> Loss: 1.30785059929 Epoch: 69 -> Test Accuracy: 53.72 [70, 60] loss: 1.159 [70, 120] loss: 1.147 [70, 180] loss: 1.123 [70, 240] loss: 1.141 [70, 300] loss: 1.126 [70, 360] loss: 1.146 Epoch: 70 -> Loss: 1.14509820938 Epoch: 70 -> Test Accuracy: 53.64 [71, 60] loss: 1.130 [71, 120] loss: 1.149 [71, 180] loss: 1.148 [71, 240] loss: 1.131 [71, 300] loss: 1.147 [71, 360] loss: 1.149 Epoch: 71 -> Loss: 1.16714203358 Epoch: 71 -> Test Accuracy: 53.67 [72, 60] loss: 1.148 [72, 120] loss: 1.151 [72, 180] loss: 1.148 [72, 240] loss: 1.147 [72, 300] loss: 1.151 [72, 360] loss: 1.129 Epoch: 72 -> Loss: 1.25139403343 Epoch: 72 -> Test Accuracy: 53.52 [73, 60] loss: 1.132 [73, 120] loss: 1.132 [73, 180] loss: 1.145 [73, 240] loss: 1.145 [73, 300] loss: 1.155 [73, 360] loss: 1.129 Epoch: 73 -> Loss: 1.00747525692 Epoch: 73 -> Test Accuracy: 53.7 [74, 60] loss: 1.147 [74, 120] loss: 1.131 [74, 180] loss: 1.149 [74, 240] loss: 1.129 [74, 300] loss: 1.148 [74, 360] loss: 1.133 Epoch: 74 -> Loss: 1.15051519871 Epoch: 74 -> Test Accuracy: 53.53 [75, 60] loss: 1.149 [75, 120] loss: 1.149 [75, 180] loss: 1.135 [75, 240] loss: 1.133 [75, 300] loss: 1.164 [75, 360] loss: 1.124 Epoch: 75 -> Loss: 1.25675094128 Epoch: 75 -> Test Accuracy: 53.54 [76, 60] loss: 1.145 [76, 120] loss: 1.133 [76, 180] loss: 1.144 [76, 240] loss: 1.135 [76, 300] loss: 1.127 [76, 360] loss: 1.135 Epoch: 76 -> Loss: 1.25496768951 Epoch: 76 -> Test Accuracy: 53.44 [77, 60] loss: 1.115 [77, 120] loss: 1.136 [77, 180] loss: 1.138 [77, 240] loss: 1.126 [77, 300] loss: 1.140 [77, 360] loss: 1.151 Epoch: 77 -> Loss: 1.1774790287 Epoch: 77 -> Test Accuracy: 53.59 [78, 60] loss: 1.157 [78, 120] loss: 1.123 [78, 180] loss: 1.119 [78, 240] loss: 1.122 [78, 300] loss: 1.118 [78, 360] loss: 1.147 Epoch: 78 -> Loss: 1.1962954998 Epoch: 78 -> Test Accuracy: 53.46 [79, 60] loss: 1.133 [79, 120] loss: 1.137 [79, 180] loss: 1.149 [79, 240] loss: 1.150 [79, 300] loss: 1.124 [79, 360] loss: 1.130 Epoch: 79 -> Loss: 1.25926578045 Epoch: 79 -> Test Accuracy: 53.64 [80, 60] loss: 1.127 [80, 120] loss: 1.124 [80, 180] loss: 1.157 [80, 240] loss: 1.112 [80, 300] loss: 1.153 [80, 360] loss: 1.130 Epoch: 80 -> Loss: 1.18208158016 Epoch: 80 -> Test Accuracy: 53.63 [81, 60] loss: 1.117 [81, 120] loss: 1.156 [81, 180] loss: 1.124 [81, 240] loss: 1.121 [81, 300] loss: 1.144 [81, 360] loss: 1.134 Epoch: 81 -> Loss: 1.16944134235 Epoch: 81 -> Test Accuracy: 53.72 [82, 60] loss: 1.122 [82, 120] loss: 1.111 [82, 180] loss: 1.142 [82, 240] loss: 1.128 [82, 300] loss: 1.127 [82, 360] loss: 1.156 Epoch: 82 -> Loss: 1.12954199314 Epoch: 82 -> Test Accuracy: 53.77 [83, 60] loss: 1.146 [83, 120] loss: 1.129 [83, 180] loss: 1.141 [83, 240] loss: 1.145 [83, 300] loss: 1.125 [83, 360] loss: 1.123 Epoch: 83 -> Loss: 1.12353086472 Epoch: 83 -> Test Accuracy: 53.68 [84, 60] loss: 1.125 [84, 120] loss: 1.138 [84, 180] loss: 1.117 [84, 240] loss: 1.119 [84, 300] loss: 1.103 [84, 360] loss: 1.155 Epoch: 84 -> Loss: 1.06623804569 Epoch: 84 -> Test Accuracy: 53.4 [85, 60] loss: 1.172 [85, 120] loss: 1.140 [85, 180] loss: 1.127 [85, 240] loss: 1.144 [85, 300] loss: 1.121 [85, 360] loss: 1.126 Epoch: 85 -> Loss: 1.21346509457 Epoch: 85 -> Test Accuracy: 53.92 [86, 60] loss: 1.119 [86, 120] loss: 1.139 [86, 180] loss: 1.129 [86, 240] loss: 1.126 [86, 300] loss: 1.136 [86, 360] loss: 1.134 Epoch: 86 -> Loss: 1.1310646534 Epoch: 86 -> Test Accuracy: 53.52 [87, 60] loss: 1.121 [87, 120] loss: 1.119 [87, 180] loss: 1.146 [87, 240] loss: 1.142 [87, 300] loss: 1.117 [87, 360] loss: 1.149 Epoch: 87 -> Loss: 1.17439734936 Epoch: 87 -> Test Accuracy: 53.7 [88, 60] loss: 1.134 [88, 120] loss: 1.140 [88, 180] loss: 1.125 [88, 240] loss: 1.131 [88, 300] loss: 1.134 [88, 360] loss: 1.128 Epoch: 88 -> Loss: 1.14962530136 Epoch: 88 -> Test Accuracy: 54.18 [89, 60] loss: 1.120 [89, 120] loss: 1.138 [89, 180] loss: 1.125 [89, 240] loss: 1.129 [89, 300] loss: 1.125 [89, 360] loss: 1.122 Epoch: 89 -> Loss: 1.06909906864 Epoch: 89 -> Test Accuracy: 53.87 [90, 60] loss: 1.138 [90, 120] loss: 1.121 [90, 180] loss: 1.117 [90, 240] loss: 1.131 [90, 300] loss: 1.153 [90, 360] loss: 1.127 Epoch: 90 -> Loss: 0.968562722206 Epoch: 90 -> Test Accuracy: 53.92 [91, 60] loss: 1.131 [91, 120] loss: 1.122 [91, 180] loss: 1.118 [91, 240] loss: 1.140 [91, 300] loss: 1.108 [91, 360] loss: 1.129 Epoch: 91 -> Loss: 1.174051404 Epoch: 91 -> Test Accuracy: 53.87 [92, 60] loss: 1.144 [92, 120] loss: 1.129 [92, 180] loss: 1.132 [92, 240] loss: 1.123 [92, 300] loss: 1.128 [92, 360] loss: 1.119 Epoch: 92 -> Loss: 1.17889094353 Epoch: 92 -> Test Accuracy: 53.6 [93, 60] loss: 1.131 [93, 120] loss: 1.130 [93, 180] loss: 1.109 [93, 240] loss: 1.120 [93, 300] loss: 1.127 [93, 360] loss: 1.142 Epoch: 93 -> Loss: 0.920753359795 Epoch: 93 -> Test Accuracy: 53.92 [94, 60] loss: 1.133 [94, 120] loss: 1.138 [94, 180] loss: 1.126 [94, 240] loss: 1.136 [94, 300] loss: 1.139 [94, 360] loss: 1.125 Epoch: 94 -> Loss: 1.15011847019 Epoch: 94 -> Test Accuracy: 53.95 [95, 60] loss: 1.127 [95, 120] loss: 1.141 [95, 180] loss: 1.115 [95, 240] loss: 1.126 [95, 300] loss: 1.118 [95, 360] loss: 1.151 Epoch: 95 -> Loss: 1.27565062046 Epoch: 95 -> Test Accuracy: 54.02 [96, 60] loss: 1.117 [96, 120] loss: 1.108 [96, 180] loss: 1.125 [96, 240] loss: 1.121 [96, 300] loss: 1.119 [96, 360] loss: 1.129 Epoch: 96 -> Loss: 1.32816052437 Epoch: 96 -> Test Accuracy: 54.13 [97, 60] loss: 1.134 [97, 120] loss: 1.155 [97, 180] loss: 1.109 [97, 240] loss: 1.127 [97, 300] loss: 1.130 [97, 360] loss: 1.130 Epoch: 97 -> Loss: 1.05178713799 Epoch: 97 -> Test Accuracy: 54.07 [98, 60] loss: 1.100 [98, 120] loss: 1.136 [98, 180] loss: 1.123 [98, 240] loss: 1.109 [98, 300] loss: 1.129 [98, 360] loss: 1.098 Epoch: 98 -> Loss: 1.00269603729 Epoch: 98 -> Test Accuracy: 54.09 [99, 60] loss: 1.130 [99, 120] loss: 1.112 [99, 180] loss: 1.125 [99, 240] loss: 1.124 [99, 300] loss: 1.121 [99, 360] loss: 1.140 Epoch: 99 -> Loss: 1.08698892593 Epoch: 99 -> Test Accuracy: 54.22 [100, 60] loss: 1.101 [100, 120] loss: 1.119 [100, 180] loss: 1.129 [100, 240] loss: 1.129 [100, 300] loss: 1.136 [100, 360] loss: 1.118 Epoch: 100 -> Loss: 1.34540462494 Epoch: 100 -> Test Accuracy: 54.07 Finished Training
In [15]:
# train ConvClassifiers on feature map of net_3block
conv_block3_loss_log, _, conv_block3_test_accuracy_log, _, _ = tr.train_all_blocks(3, 10, [0.1, 0.02, 0.004, 0.0008],
[35, 70, 85, 100], 0.9, 5e-4, net_block3, criterion, trainloader, None, testloader, use_ConvClassifier=True)
[1, 60] loss: 1.371 [1, 120] loss: 1.023 [1, 180] loss: 0.913 [1, 240] loss: 0.887 [1, 300] loss: 0.824 [1, 360] loss: 0.783 Epoch: 1 -> Loss: 0.617035627365 Epoch: 1 -> Test Accuracy: 70.87 [2, 60] loss: 0.747 [2, 120] loss: 0.703 [2, 180] loss: 0.689 [2, 240] loss: 0.676 [2, 300] loss: 0.673 [2, 360] loss: 0.660 Epoch: 2 -> Loss: 0.749189436436 Epoch: 2 -> Test Accuracy: 74.43 [3, 60] loss: 0.627 [3, 120] loss: 0.612 [3, 180] loss: 0.619 [3, 240] loss: 0.611 [3, 300] loss: 0.628 [3, 360] loss: 0.587 Epoch: 3 -> Loss: 0.666461706161 Epoch: 3 -> Test Accuracy: 77.75 [4, 60] loss: 0.566 [4, 120] loss: 0.570 [4, 180] loss: 0.566 [4, 240] loss: 0.556 [4, 300] loss: 0.553 [4, 360] loss: 0.576 Epoch: 4 -> Loss: 0.634184956551 Epoch: 4 -> Test Accuracy: 77.39 [5, 60] loss: 0.521 [5, 120] loss: 0.541 [5, 180] loss: 0.559 [5, 240] loss: 0.535 [5, 300] loss: 0.520 [5, 360] loss: 0.528 Epoch: 5 -> Loss: 0.685723721981 Epoch: 5 -> Test Accuracy: 78.43 [6, 60] loss: 0.509 [6, 120] loss: 0.506 [6, 180] loss: 0.529 [6, 240] loss: 0.504 [6, 300] loss: 0.516 [6, 360] loss: 0.521 Epoch: 6 -> Loss: 0.714321732521 Epoch: 6 -> Test Accuracy: 77.98 [7, 60] loss: 0.514 [7, 120] loss: 0.508 [7, 180] loss: 0.497 [7, 240] loss: 0.484 [7, 300] loss: 0.491 [7, 360] loss: 0.506 Epoch: 7 -> Loss: 0.479330956936 Epoch: 7 -> Test Accuracy: 79.51 [8, 60] loss: 0.460 [8, 120] loss: 0.470 [8, 180] loss: 0.499 [8, 240] loss: 0.493 [8, 300] loss: 0.505 [8, 360] loss: 0.483 Epoch: 8 -> Loss: 0.708630979061 Epoch: 8 -> Test Accuracy: 79.96 [9, 60] loss: 0.445 [9, 120] loss: 0.464 [9, 180] loss: 0.486 [9, 240] loss: 0.468 [9, 300] loss: 0.493 [9, 360] loss: 0.472 Epoch: 9 -> Loss: 0.668149590492 Epoch: 9 -> Test Accuracy: 80.27 [10, 60] loss: 0.449 [10, 120] loss: 0.466 [10, 180] loss: 0.485 [10, 240] loss: 0.492 [10, 300] loss: 0.481 [10, 360] loss: 0.469 Epoch: 10 -> Loss: 0.470273196697 Epoch: 10 -> Test Accuracy: 80.33 [11, 60] loss: 0.439 [11, 120] loss: 0.468 [11, 180] loss: 0.455 [11, 240] loss: 0.461 [11, 300] loss: 0.471 [11, 360] loss: 0.454 Epoch: 11 -> Loss: 0.494262129068 Epoch: 11 -> Test Accuracy: 80.4 [12, 60] loss: 0.447 [12, 120] loss: 0.437 [12, 180] loss: 0.464 [12, 240] loss: 0.447 [12, 300] loss: 0.460 [12, 360] loss: 0.468 Epoch: 12 -> Loss: 0.424614042044 Epoch: 12 -> Test Accuracy: 80.15 [13, 60] loss: 0.431 [13, 120] loss: 0.437 [13, 180] loss: 0.458 [13, 240] loss: 0.448 [13, 300] loss: 0.456 [13, 360] loss: 0.459 Epoch: 13 -> Loss: 0.420501470566 Epoch: 13 -> Test Accuracy: 81.55 [14, 60] loss: 0.425 [14, 120] loss: 0.447 [14, 180] loss: 0.428 [14, 240] loss: 0.441 [14, 300] loss: 0.462 [14, 360] loss: 0.445 Epoch: 14 -> Loss: 0.493283182383 Epoch: 14 -> Test Accuracy: 80.95 [15, 60] loss: 0.415 [15, 120] loss: 0.409 [15, 180] loss: 0.452 [15, 240] loss: 0.435 [15, 300] loss: 0.458 [15, 360] loss: 0.453 Epoch: 15 -> Loss: 0.687053024769 Epoch: 15 -> Test Accuracy: 81.13 [16, 60] loss: 0.426 [16, 120] loss: 0.402 [16, 180] loss: 0.449 [16, 240] loss: 0.436 [16, 300] loss: 0.451 [16, 360] loss: 0.433 Epoch: 16 -> Loss: 0.525857210159 Epoch: 16 -> Test Accuracy: 82.14 [17, 60] loss: 0.412 [17, 120] loss: 0.425 [17, 180] loss: 0.454 [17, 240] loss: 0.423 [17, 300] loss: 0.433 [17, 360] loss: 0.444 Epoch: 17 -> Loss: 0.48986697197 Epoch: 17 -> Test Accuracy: 81.2 [18, 60] loss: 0.416 [18, 120] loss: 0.417 [18, 180] loss: 0.447 [18, 240] loss: 0.436 [18, 300] loss: 0.441 [18, 360] loss: 0.430 Epoch: 18 -> Loss: 0.359789043665 Epoch: 18 -> Test Accuracy: 81.58 [19, 60] loss: 0.400 [19, 120] loss: 0.415 [19, 180] loss: 0.427 [19, 240] loss: 0.431 [19, 300] loss: 0.448 [19, 360] loss: 0.434 Epoch: 19 -> Loss: 0.476075470448 Epoch: 19 -> Test Accuracy: 81.62 [20, 60] loss: 0.408 [20, 120] loss: 0.411 [20, 180] loss: 0.412 [20, 240] loss: 0.436 [20, 300] loss: 0.436 [20, 360] loss: 0.406 Epoch: 20 -> Loss: 0.665274560452 Epoch: 20 -> Test Accuracy: 80.63 [21, 60] loss: 0.413 [21, 120] loss: 0.413 [21, 180] loss: 0.412 [21, 240] loss: 0.431 [21, 300] loss: 0.426 [21, 360] loss: 0.450 Epoch: 21 -> Loss: 0.453517138958 Epoch: 21 -> Test Accuracy: 82.28 [22, 60] loss: 0.392 [22, 120] loss: 0.415 [22, 180] loss: 0.409 [22, 240] loss: 0.427 [22, 300] loss: 0.420 [22, 360] loss: 0.428 Epoch: 22 -> Loss: 0.365581929684 Epoch: 22 -> Test Accuracy: 79.74 [23, 60] loss: 0.400 [23, 120] loss: 0.415 [23, 180] loss: 0.434 [23, 240] loss: 0.427 [23, 300] loss: 0.427 [23, 360] loss: 0.436 Epoch: 23 -> Loss: 0.307698786259 Epoch: 23 -> Test Accuracy: 80.45 [24, 60] loss: 0.403 [24, 120] loss: 0.411 [24, 180] loss: 0.403 [24, 240] loss: 0.418 [24, 300] loss: 0.409 [24, 360] loss: 0.424 Epoch: 24 -> Loss: 0.541094362736 Epoch: 24 -> Test Accuracy: 82.16 [25, 60] loss: 0.412 [25, 120] loss: 0.401 [25, 180] loss: 0.421 [25, 240] loss: 0.435 [25, 300] loss: 0.417 [25, 360] loss: 0.424 Epoch: 25 -> Loss: 0.558412909508 Epoch: 25 -> Test Accuracy: 81.44 [26, 60] loss: 0.409 [26, 120] loss: 0.408 [26, 180] loss: 0.403 [26, 240] loss: 0.423 [26, 300] loss: 0.421 [26, 360] loss: 0.410 Epoch: 26 -> Loss: 0.406867265701 Epoch: 26 -> Test Accuracy: 81.33 [27, 60] loss: 0.391 [27, 120] loss: 0.404 [27, 180] loss: 0.427 [27, 240] loss: 0.406 [27, 300] loss: 0.413 [27, 360] loss: 0.427 Epoch: 27 -> Loss: 0.352363586426 Epoch: 27 -> Test Accuracy: 81.67 [28, 60] loss: 0.389 [28, 120] loss: 0.404 [28, 180] loss: 0.415 [28, 240] loss: 0.413 [28, 300] loss: 0.414 [28, 360] loss: 0.421 Epoch: 28 -> Loss: 0.485638141632 Epoch: 28 -> Test Accuracy: 81.59 [29, 60] loss: 0.389 [29, 120] loss: 0.400 [29, 180] loss: 0.399 [29, 240] loss: 0.424 [29, 300] loss: 0.408 [29, 360] loss: 0.434 Epoch: 29 -> Loss: 0.435752779245 Epoch: 29 -> Test Accuracy: 82.08 [30, 60] loss: 0.392 [30, 120] loss: 0.401 [30, 180] loss: 0.392 [30, 240] loss: 0.415 [30, 300] loss: 0.417 [30, 360] loss: 0.420 Epoch: 30 -> Loss: 0.471777528524 Epoch: 30 -> Test Accuracy: 81.43 [31, 60] loss: 0.414 [31, 120] loss: 0.396 [31, 180] loss: 0.409 [31, 240] loss: 0.416 [31, 300] loss: 0.404 [31, 360] loss: 0.416 Epoch: 31 -> Loss: 0.586894094944 Epoch: 31 -> Test Accuracy: 81.65 [32, 60] loss: 0.394 [32, 120] loss: 0.389 [32, 180] loss: 0.424 [32, 240] loss: 0.409 [32, 300] loss: 0.413 [32, 360] loss: 0.408 Epoch: 32 -> Loss: 0.67364937067 Epoch: 32 -> Test Accuracy: 81.55 [33, 60] loss: 0.387 [33, 120] loss: 0.389 [33, 180] loss: 0.415 [33, 240] loss: 0.418 [33, 300] loss: 0.405 [33, 360] loss: 0.398 Epoch: 33 -> Loss: 0.429774224758 Epoch: 33 -> Test Accuracy: 81.84 [34, 60] loss: 0.385 [34, 120] loss: 0.392 [34, 180] loss: 0.420 [34, 240] loss: 0.404 [34, 300] loss: 0.402 [34, 360] loss: 0.434 Epoch: 34 -> Loss: 0.351295530796 Epoch: 34 -> Test Accuracy: 82.5 [35, 60] loss: 0.369 [35, 120] loss: 0.407 [35, 180] loss: 0.403 [35, 240] loss: 0.411 [35, 300] loss: 0.416 [35, 360] loss: 0.411 Epoch: 35 -> Loss: 0.46245008707 Epoch: 35 -> Test Accuracy: 81.89 [36, 60] loss: 0.307 [36, 120] loss: 0.282 [36, 180] loss: 0.271 [36, 240] loss: 0.284 [36, 300] loss: 0.279 [36, 360] loss: 0.271 Epoch: 36 -> Loss: 0.511943459511 Epoch: 36 -> Test Accuracy: 85.62 [37, 60] loss: 0.255 [37, 120] loss: 0.250 [37, 180] loss: 0.244 [37, 240] loss: 0.248 [37, 300] loss: 0.248 [37, 360] loss: 0.251 Epoch: 37 -> Loss: 0.307639151812 Epoch: 37 -> Test Accuracy: 86.44 [38, 60] loss: 0.231 [38, 120] loss: 0.237 [38, 180] loss: 0.247 [38, 240] loss: 0.225 [38, 300] loss: 0.230 [38, 360] loss: 0.252 Epoch: 38 -> Loss: 0.353818029165 Epoch: 38 -> Test Accuracy: 85.76 [39, 60] loss: 0.223 [39, 120] loss: 0.223 [39, 180] loss: 0.219 [39, 240] loss: 0.231 [39, 300] loss: 0.237 [39, 360] loss: 0.242 Epoch: 39 -> Loss: 0.209128811955 Epoch: 39 -> Test Accuracy: 85.6 [40, 60] loss: 0.212 [40, 120] loss: 0.229 [40, 180] loss: 0.229 [40, 240] loss: 0.229 [40, 300] loss: 0.215 [40, 360] loss: 0.229 Epoch: 40 -> Loss: 0.214634135365 Epoch: 40 -> Test Accuracy: 86.36 [41, 60] loss: 0.203 [41, 120] loss: 0.227 [41, 180] loss: 0.219 [41, 240] loss: 0.222 [41, 300] loss: 0.232 [41, 360] loss: 0.238 Epoch: 41 -> Loss: 0.193170338869 Epoch: 41 -> Test Accuracy: 85.79 [42, 60] loss: 0.210 [42, 120] loss: 0.210 [42, 180] loss: 0.213 [42, 240] loss: 0.220 [42, 300] loss: 0.226 [42, 360] loss: 0.229 Epoch: 42 -> Loss: 0.367429107428 Epoch: 42 -> Test Accuracy: 85.27 [43, 60] loss: 0.208 [43, 120] loss: 0.212 [43, 180] loss: 0.223 [43, 240] loss: 0.226 [43, 300] loss: 0.225 [43, 360] loss: 0.230 Epoch: 43 -> Loss: 0.246880248189 Epoch: 43 -> Test Accuracy: 86.26 [44, 60] loss: 0.211 [44, 120] loss: 0.199 [44, 180] loss: 0.218 [44, 240] loss: 0.218 [44, 300] loss: 0.224 [44, 360] loss: 0.239 Epoch: 44 -> Loss: 0.308743089437 Epoch: 44 -> Test Accuracy: 85.1 [45, 60] loss: 0.207 [45, 120] loss: 0.211 [45, 180] loss: 0.216 [45, 240] loss: 0.227 [45, 300] loss: 0.221 [45, 360] loss: 0.218 Epoch: 45 -> Loss: 0.193989947438 Epoch: 45 -> Test Accuracy: 84.7 [46, 60] loss: 0.211 [46, 120] loss: 0.204 [46, 180] loss: 0.213 [46, 240] loss: 0.218 [46, 300] loss: 0.221 [46, 360] loss: 0.237 Epoch: 46 -> Loss: 0.173785120249 Epoch: 46 -> Test Accuracy: 85.62 [47, 60] loss: 0.209 [47, 120] loss: 0.213 [47, 180] loss: 0.213 [47, 240] loss: 0.232 [47, 300] loss: 0.227 [47, 360] loss: 0.235 Epoch: 47 -> Loss: 0.215796589851 Epoch: 47 -> Test Accuracy: 85.7 [48, 60] loss: 0.197 [48, 120] loss: 0.203 [48, 180] loss: 0.215 [48, 240] loss: 0.224 [48, 300] loss: 0.228 [48, 360] loss: 0.239 Epoch: 48 -> Loss: 0.272399038076 Epoch: 48 -> Test Accuracy: 84.95 [49, 60] loss: 0.205 [49, 120] loss: 0.206 [49, 180] loss: 0.222 [49, 240] loss: 0.219 [49, 300] loss: 0.231 [49, 360] loss: 0.235 Epoch: 49 -> Loss: 0.123649761081 Epoch: 49 -> Test Accuracy: 84.89 [50, 60] loss: 0.205 [50, 120] loss: 0.210 [50, 180] loss: 0.224 [50, 240] loss: 0.219 [50, 300] loss: 0.220 [50, 360] loss: 0.233 Epoch: 50 -> Loss: 0.227295681834 Epoch: 50 -> Test Accuracy: 85.24 [51, 60] loss: 0.205 [51, 120] loss: 0.211 [51, 180] loss: 0.209 [51, 240] loss: 0.217 [51, 300] loss: 0.221 [51, 360] loss: 0.226 Epoch: 51 -> Loss: 0.302188068628 Epoch: 51 -> Test Accuracy: 85.78 [52, 60] loss: 0.203 [52, 120] loss: 0.195 [52, 180] loss: 0.214 [52, 240] loss: 0.216 [52, 300] loss: 0.229 [52, 360] loss: 0.228 Epoch: 52 -> Loss: 0.26025018096 Epoch: 52 -> Test Accuracy: 85.12 [53, 60] loss: 0.198 [53, 120] loss: 0.201 [53, 180] loss: 0.228 [53, 240] loss: 0.210 [53, 300] loss: 0.228 [53, 360] loss: 0.236 Epoch: 53 -> Loss: 0.345288306475 Epoch: 53 -> Test Accuracy: 85.29 [54, 60] loss: 0.199 [54, 120] loss: 0.199 [54, 180] loss: 0.202 [54, 240] loss: 0.231 [54, 300] loss: 0.223 [54, 360] loss: 0.232 Epoch: 54 -> Loss: 0.20453453064 Epoch: 54 -> Test Accuracy: 85.02 [55, 60] loss: 0.212 [55, 120] loss: 0.215 [55, 180] loss: 0.216 [55, 240] loss: 0.211 [55, 300] loss: 0.222 [55, 360] loss: 0.232 Epoch: 55 -> Loss: 0.144998937845 Epoch: 55 -> Test Accuracy: 84.86 [56, 60] loss: 0.202 [56, 120] loss: 0.198 [56, 180] loss: 0.212 [56, 240] loss: 0.219 [56, 300] loss: 0.236 [56, 360] loss: 0.223 Epoch: 56 -> Loss: 0.226112693548 Epoch: 56 -> Test Accuracy: 84.92 [57, 60] loss: 0.195 [57, 120] loss: 0.208 [57, 180] loss: 0.218 [57, 240] loss: 0.213 [57, 300] loss: 0.221 [57, 360] loss: 0.228 Epoch: 57 -> Loss: 0.235637187958 Epoch: 57 -> Test Accuracy: 84.86 [58, 60] loss: 0.198 [58, 120] loss: 0.211 [58, 180] loss: 0.207 [58, 240] loss: 0.219 [58, 300] loss: 0.217 [58, 360] loss: 0.224 Epoch: 58 -> Loss: 0.234650462866 Epoch: 58 -> Test Accuracy: 84.77 [59, 60] loss: 0.199 [59, 120] loss: 0.198 [59, 180] loss: 0.211 [59, 240] loss: 0.218 [59, 300] loss: 0.219 [59, 360] loss: 0.233 Epoch: 59 -> Loss: 0.1255761832 Epoch: 59 -> Test Accuracy: 84.43 [60, 60] loss: 0.197 [60, 120] loss: 0.209 [60, 180] loss: 0.204 [60, 240] loss: 0.217 [60, 300] loss: 0.216 [60, 360] loss: 0.220 Epoch: 60 -> Loss: 0.277145057917 Epoch: 60 -> Test Accuracy: 85.29 [61, 60] loss: 0.188 [61, 120] loss: 0.190 [61, 180] loss: 0.207 [61, 240] loss: 0.213 [61, 300] loss: 0.216 [61, 360] loss: 0.226 Epoch: 61 -> Loss: 0.139841303229 Epoch: 61 -> Test Accuracy: 85.05 [62, 60] loss: 0.196 [62, 120] loss: 0.213 [62, 180] loss: 0.198 [62, 240] loss: 0.217 [62, 300] loss: 0.217 [62, 360] loss: 0.228 Epoch: 62 -> Loss: 0.245643734932 Epoch: 62 -> Test Accuracy: 85.33 [63, 60] loss: 0.193 [63, 120] loss: 0.202 [63, 180] loss: 0.208 [63, 240] loss: 0.200 [63, 300] loss: 0.211 [63, 360] loss: 0.232 Epoch: 63 -> Loss: 0.336574912071 Epoch: 63 -> Test Accuracy: 85.29 [64, 60] loss: 0.205 [64, 120] loss: 0.202 [64, 180] loss: 0.214 [64, 240] loss: 0.210 [64, 300] loss: 0.210 [64, 360] loss: 0.222 Epoch: 64 -> Loss: 0.157260462642 Epoch: 64 -> Test Accuracy: 85.26 [65, 60] loss: 0.196 [65, 120] loss: 0.200 [65, 180] loss: 0.195 [65, 240] loss: 0.204 [65, 300] loss: 0.212 [65, 360] loss: 0.219 Epoch: 65 -> Loss: 0.289101332426 Epoch: 65 -> Test Accuracy: 84.37 [66, 60] loss: 0.190 [66, 120] loss: 0.204 [66, 180] loss: 0.220 [66, 240] loss: 0.210 [66, 300] loss: 0.215 [66, 360] loss: 0.218 Epoch: 66 -> Loss: 0.192337989807 Epoch: 66 -> Test Accuracy: 84.95 [67, 60] loss: 0.195 [67, 120] loss: 0.191 [67, 180] loss: 0.213 [67, 240] loss: 0.207 [67, 300] loss: 0.223 [67, 360] loss: 0.218 Epoch: 67 -> Loss: 0.191490486264 Epoch: 67 -> Test Accuracy: 84.88 [68, 60] loss: 0.191 [68, 120] loss: 0.189 [68, 180] loss: 0.210 [68, 240] loss: 0.216 [68, 300] loss: 0.226 [68, 360] loss: 0.222 Epoch: 68 -> Loss: 0.217858999968 Epoch: 68 -> Test Accuracy: 83.97 [69, 60] loss: 0.194 [69, 120] loss: 0.185 [69, 180] loss: 0.195 [69, 240] loss: 0.211 [69, 300] loss: 0.209 [69, 360] loss: 0.219 Epoch: 69 -> Loss: 0.202600002289 Epoch: 69 -> Test Accuracy: 84.85 [70, 60] loss: 0.196 [70, 120] loss: 0.191 [70, 180] loss: 0.199 [70, 240] loss: 0.212 [70, 300] loss: 0.213 [70, 360] loss: 0.214 Epoch: 70 -> Loss: 0.217269584537 Epoch: 70 -> Test Accuracy: 85.36 [71, 60] loss: 0.173 [71, 120] loss: 0.144 [71, 180] loss: 0.140 [71, 240] loss: 0.136 [71, 300] loss: 0.140 [71, 360] loss: 0.133 Epoch: 71 -> Loss: 0.0861133784056 Epoch: 71 -> Test Accuracy: 86.88 [72, 60] loss: 0.121 [72, 120] loss: 0.121 [72, 180] loss: 0.122 [72, 240] loss: 0.120 [72, 300] loss: 0.128 [72, 360] loss: 0.124 Epoch: 72 -> Loss: 0.152239322662 Epoch: 72 -> Test Accuracy: 86.78 [73, 60] loss: 0.124 [73, 120] loss: 0.119 [73, 180] loss: 0.115 [73, 240] loss: 0.116 [73, 300] loss: 0.124 [73, 360] loss: 0.125 Epoch: 73 -> Loss: 0.24748647213 Epoch: 73 -> Test Accuracy: 86.97 [74, 60] loss: 0.112 [74, 120] loss: 0.109 [74, 180] loss: 0.112 [74, 240] loss: 0.122 [74, 300] loss: 0.112 [74, 360] loss: 0.122 Epoch: 74 -> Loss: 0.152580738068 Epoch: 74 -> Test Accuracy: 86.8 [75, 60] loss: 0.109 [75, 120] loss: 0.115 [75, 180] loss: 0.109 [75, 240] loss: 0.119 [75, 300] loss: 0.110 [75, 360] loss: 0.109 Epoch: 75 -> Loss: 0.0667220279574 Epoch: 75 -> Test Accuracy: 86.88 [76, 60] loss: 0.103 [76, 120] loss: 0.104 [76, 180] loss: 0.103 [76, 240] loss: 0.110 [76, 300] loss: 0.111 [76, 360] loss: 0.119 Epoch: 76 -> Loss: 0.104675829411 Epoch: 76 -> Test Accuracy: 87.01 [77, 60] loss: 0.105 [77, 120] loss: 0.109 [77, 180] loss: 0.105 [77, 240] loss: 0.107 [77, 300] loss: 0.109 [77, 360] loss: 0.112 Epoch: 77 -> Loss: 0.0587496869266 Epoch: 77 -> Test Accuracy: 86.84 [78, 60] loss: 0.104 [78, 120] loss: 0.103 [78, 180] loss: 0.107 [78, 240] loss: 0.102 [78, 300] loss: 0.100 [78, 360] loss: 0.104 Epoch: 78 -> Loss: 0.119968272746 Epoch: 78 -> Test Accuracy: 87.15 [79, 60] loss: 0.098 [79, 120] loss: 0.103 [79, 180] loss: 0.099 [79, 240] loss: 0.100 [79, 300] loss: 0.107 [79, 360] loss: 0.110 Epoch: 79 -> Loss: 0.148834779859 Epoch: 79 -> Test Accuracy: 86.81 [80, 60] loss: 0.101 [80, 120] loss: 0.109 [80, 180] loss: 0.102 [80, 240] loss: 0.096 [80, 300] loss: 0.102 [80, 360] loss: 0.108 Epoch: 80 -> Loss: 0.112207576632 Epoch: 80 -> Test Accuracy: 86.9 [81, 60] loss: 0.092 [81, 120] loss: 0.103 [81, 180] loss: 0.103 [81, 240] loss: 0.101 [81, 300] loss: 0.101 [81, 360] loss: 0.100 Epoch: 81 -> Loss: 0.0783820748329 Epoch: 81 -> Test Accuracy: 86.72 [82, 60] loss: 0.097 [82, 120] loss: 0.098 [82, 180] loss: 0.098 [82, 240] loss: 0.094 [82, 300] loss: 0.096 [82, 360] loss: 0.096 Epoch: 82 -> Loss: 0.147284641862 Epoch: 82 -> Test Accuracy: 86.91 [83, 60] loss: 0.097 [83, 120] loss: 0.100 [83, 180] loss: 0.095 [83, 240] loss: 0.105 [83, 300] loss: 0.100 [83, 360] loss: 0.096 Epoch: 83 -> Loss: 0.0895229056478 Epoch: 83 -> Test Accuracy: 86.76 [84, 60] loss: 0.093 [84, 120] loss: 0.098 [84, 180] loss: 0.095 [84, 240] loss: 0.098 [84, 300] loss: 0.099 [84, 360] loss: 0.100 Epoch: 84 -> Loss: 0.173047661781 Epoch: 84 -> Test Accuracy: 86.94 [85, 60] loss: 0.092 [85, 120] loss: 0.090 [85, 180] loss: 0.095 [85, 240] loss: 0.098 [85, 300] loss: 0.094 [85, 360] loss: 0.098 Epoch: 85 -> Loss: 0.0852631404996 Epoch: 85 -> Test Accuracy: 87.11 [86, 60] loss: 0.092 [86, 120] loss: 0.083 [86, 180] loss: 0.082 [86, 240] loss: 0.079 [86, 300] loss: 0.082 [86, 360] loss: 0.087 Epoch: 86 -> Loss: 0.117199338973 Epoch: 86 -> Test Accuracy: 87.23 [87, 60] loss: 0.080 [87, 120] loss: 0.077 [87, 180] loss: 0.080 [87, 240] loss: 0.082 [87, 300] loss: 0.080 [87, 360] loss: 0.081 Epoch: 87 -> Loss: 0.0730190724134 Epoch: 87 -> Test Accuracy: 87.05 [88, 60] loss: 0.080 [88, 120] loss: 0.079 [88, 180] loss: 0.082 [88, 240] loss: 0.077 [88, 300] loss: 0.079 [88, 360] loss: 0.077 Epoch: 88 -> Loss: 0.0719773620367 Epoch: 88 -> Test Accuracy: 87.09 [89, 60] loss: 0.079 [89, 120] loss: 0.081 [89, 180] loss: 0.074 [89, 240] loss: 0.079 [89, 300] loss: 0.074 [89, 360] loss: 0.078 Epoch: 89 -> Loss: 0.0586576089263 Epoch: 89 -> Test Accuracy: 87.1 [90, 60] loss: 0.076 [90, 120] loss: 0.082 [90, 180] loss: 0.081 [90, 240] loss: 0.081 [90, 300] loss: 0.075 [90, 360] loss: 0.080 Epoch: 90 -> Loss: 0.068951241672 Epoch: 90 -> Test Accuracy: 87.01 [91, 60] loss: 0.077 [91, 120] loss: 0.075 [91, 180] loss: 0.078 [91, 240] loss: 0.079 [91, 300] loss: 0.082 [91, 360] loss: 0.078 Epoch: 91 -> Loss: 0.062657892704 Epoch: 91 -> Test Accuracy: 87.12 [92, 60] loss: 0.079 [92, 120] loss: 0.078 [92, 180] loss: 0.076 [92, 240] loss: 0.078 [92, 300] loss: 0.080 [92, 360] loss: 0.080 Epoch: 92 -> Loss: 0.0748406276107 Epoch: 92 -> Test Accuracy: 86.95 [93, 60] loss: 0.071 [93, 120] loss: 0.076 [93, 180] loss: 0.076 [93, 240] loss: 0.073 [93, 300] loss: 0.082 [93, 360] loss: 0.079 Epoch: 93 -> Loss: 0.0712976232171 Epoch: 93 -> Test Accuracy: 87.08 [94, 60] loss: 0.073 [94, 120] loss: 0.078 [94, 180] loss: 0.079 [94, 240] loss: 0.077 [94, 300] loss: 0.078 [94, 360] loss: 0.077 Epoch: 94 -> Loss: 0.136106818914 Epoch: 94 -> Test Accuracy: 87.23 [95, 60] loss: 0.077 [95, 120] loss: 0.079 [95, 180] loss: 0.076 [95, 240] loss: 0.080 [95, 300] loss: 0.079 [95, 360] loss: 0.075 Epoch: 95 -> Loss: 0.0827887803316 Epoch: 95 -> Test Accuracy: 86.9 [96, 60] loss: 0.075 [96, 120] loss: 0.078 [96, 180] loss: 0.075 [96, 240] loss: 0.078 [96, 300] loss: 0.077 [96, 360] loss: 0.075 Epoch: 96 -> Loss: 0.074126958847 Epoch: 96 -> Test Accuracy: 87.09 [97, 60] loss: 0.077 [97, 120] loss: 0.073 [97, 180] loss: 0.077 [97, 240] loss: 0.075 [97, 300] loss: 0.077 [97, 360] loss: 0.078 Epoch: 97 -> Loss: 0.123150423169 Epoch: 97 -> Test Accuracy: 86.99 [98, 60] loss: 0.076 [98, 120] loss: 0.077 [98, 180] loss: 0.076 [98, 240] loss: 0.071 [98, 300] loss: 0.074 [98, 360] loss: 0.079 Epoch: 98 -> Loss: 0.0828704237938 Epoch: 98 -> Test Accuracy: 87.05 [99, 60] loss: 0.074 [99, 120] loss: 0.075 [99, 180] loss: 0.071 [99, 240] loss: 0.077 [99, 300] loss: 0.078 [99, 360] loss: 0.077 Epoch: 99 -> Loss: 0.0612016692758 Epoch: 99 -> Test Accuracy: 87.13 [100, 60] loss: 0.069 [100, 120] loss: 0.075 [100, 180] loss: 0.073 [100, 240] loss: 0.074 [100, 300] loss: 0.074 [100, 360] loss: 0.075 Epoch: 100 -> Loss: 0.038023866713 Epoch: 100 -> Test Accuracy: 86.85 Finished Training [1, 60] loss: 0.899 [1, 120] loss: 0.623 [1, 180] loss: 0.573 [1, 240] loss: 0.570 [1, 300] loss: 0.516 [1, 360] loss: 0.494 Epoch: 1 -> Loss: 0.527269244194 Epoch: 1 -> Test Accuracy: 81.08 [2, 60] loss: 0.455 [2, 120] loss: 0.454 [2, 180] loss: 0.448 [2, 240] loss: 0.427 [2, 300] loss: 0.435 [2, 360] loss: 0.446 Epoch: 2 -> Loss: 0.406312793493 Epoch: 2 -> Test Accuracy: 83.37 [3, 60] loss: 0.394 [3, 120] loss: 0.399 [3, 180] loss: 0.400 [3, 240] loss: 0.399 [3, 300] loss: 0.408 [3, 360] loss: 0.392 Epoch: 3 -> Loss: 0.339932471514 Epoch: 3 -> Test Accuracy: 83.35 [4, 60] loss: 0.363 [4, 120] loss: 0.368 [4, 180] loss: 0.360 [4, 240] loss: 0.381 [4, 300] loss: 0.382 [4, 360] loss: 0.384 Epoch: 4 -> Loss: 0.269151031971 Epoch: 4 -> Test Accuracy: 84.44 [5, 60] loss: 0.338 [5, 120] loss: 0.354 [5, 180] loss: 0.350 [5, 240] loss: 0.341 [5, 300] loss: 0.366 [5, 360] loss: 0.356 Epoch: 5 -> Loss: 0.325666487217 Epoch: 5 -> Test Accuracy: 84.63 [6, 60] loss: 0.314 [6, 120] loss: 0.327 [6, 180] loss: 0.335 [6, 240] loss: 0.342 [6, 300] loss: 0.347 [6, 360] loss: 0.356 Epoch: 6 -> Loss: 0.313348770142 Epoch: 6 -> Test Accuracy: 83.96 [7, 60] loss: 0.319 [7, 120] loss: 0.331 [7, 180] loss: 0.324 [7, 240] loss: 0.321 [7, 300] loss: 0.341 [7, 360] loss: 0.336 Epoch: 7 -> Loss: 0.443416684866 Epoch: 7 -> Test Accuracy: 83.39 [8, 60] loss: 0.300 [8, 120] loss: 0.310 [8, 180] loss: 0.317 [8, 240] loss: 0.321 [8, 300] loss: 0.342 [8, 360] loss: 0.336 Epoch: 8 -> Loss: 0.214324861765 Epoch: 8 -> Test Accuracy: 85.53 [9, 60] loss: 0.305 [9, 120] loss: 0.302 [9, 180] loss: 0.305 [9, 240] loss: 0.325 [9, 300] loss: 0.327 [9, 360] loss: 0.315 Epoch: 9 -> Loss: 0.455975621939 Epoch: 9 -> Test Accuracy: 85.41 [10, 60] loss: 0.276 [10, 120] loss: 0.294 [10, 180] loss: 0.309 [10, 240] loss: 0.295 [10, 300] loss: 0.325 [10, 360] loss: 0.320 Epoch: 10 -> Loss: 0.491630464792 Epoch: 10 -> Test Accuracy: 85.57 [11, 60] loss: 0.271 [11, 120] loss: 0.285 [11, 180] loss: 0.289 [11, 240] loss: 0.304 [11, 300] loss: 0.339 [11, 360] loss: 0.330 Epoch: 11 -> Loss: 0.320974588394 Epoch: 11 -> Test Accuracy: 86.18 [12, 60] loss: 0.280 [12, 120] loss: 0.280 [12, 180] loss: 0.281 [12, 240] loss: 0.292 [12, 300] loss: 0.298 [12, 360] loss: 0.311 Epoch: 12 -> Loss: 0.311203598976 Epoch: 12 -> Test Accuracy: 85.38 [13, 60] loss: 0.266 [13, 120] loss: 0.289 [13, 180] loss: 0.293 [13, 240] loss: 0.303 [13, 300] loss: 0.298 [13, 360] loss: 0.304 Epoch: 13 -> Loss: 0.38963535428 Epoch: 13 -> Test Accuracy: 85.49 [14, 60] loss: 0.279 [14, 120] loss: 0.287 [14, 180] loss: 0.293 [14, 240] loss: 0.276 [14, 300] loss: 0.309 [14, 360] loss: 0.304 Epoch: 14 -> Loss: 0.336613625288 Epoch: 14 -> Test Accuracy: 85.05 [15, 60] loss: 0.260 [15, 120] loss: 0.287 [15, 180] loss: 0.293 [15, 240] loss: 0.281 [15, 300] loss: 0.278 [15, 360] loss: 0.284 Epoch: 15 -> Loss: 0.39779239893 Epoch: 15 -> Test Accuracy: 84.99 [16, 60] loss: 0.274 [16, 120] loss: 0.273 [16, 180] loss: 0.269 [16, 240] loss: 0.287 [16, 300] loss: 0.298 [16, 360] loss: 0.305 Epoch: 16 -> Loss: 0.339841604233 Epoch: 16 -> Test Accuracy: 85.91 [17, 60] loss: 0.256 [17, 120] loss: 0.273 [17, 180] loss: 0.294 [17, 240] loss: 0.283 [17, 300] loss: 0.304 [17, 360] loss: 0.299 Epoch: 17 -> Loss: 0.359031558037 Epoch: 17 -> Test Accuracy: 85.61 [18, 60] loss: 0.258 [18, 120] loss: 0.259 [18, 180] loss: 0.273 [18, 240] loss: 0.285 [18, 300] loss: 0.288 [18, 360] loss: 0.306 Epoch: 18 -> Loss: 0.35723093152 Epoch: 18 -> Test Accuracy: 85.08 [19, 60] loss: 0.284 [19, 120] loss: 0.280 [19, 180] loss: 0.278 [19, 240] loss: 0.276 [19, 300] loss: 0.277 [19, 360] loss: 0.288 Epoch: 19 -> Loss: 0.2690769732 Epoch: 19 -> Test Accuracy: 85.38 [20, 60] loss: 0.261 [20, 120] loss: 0.256 [20, 180] loss: 0.284 [20, 240] loss: 0.284 [20, 300] loss: 0.278 [20, 360] loss: 0.301 Epoch: 20 -> Loss: 0.276969313622 Epoch: 20 -> Test Accuracy: 86.02 [21, 60] loss: 0.264 [21, 120] loss: 0.265 [21, 180] loss: 0.252 [21, 240] loss: 0.278 [21, 300] loss: 0.289 [21, 360] loss: 0.303 Epoch: 21 -> Loss: 0.314692467451 Epoch: 21 -> Test Accuracy: 85.43 [22, 60] loss: 0.257 [22, 120] loss: 0.264 [22, 180] loss: 0.274 [22, 240] loss: 0.277 [22, 300] loss: 0.281 [22, 360] loss: 0.284 Epoch: 22 -> Loss: 0.241112902761 Epoch: 22 -> Test Accuracy: 85.38 [23, 60] loss: 0.250 [23, 120] loss: 0.259 [23, 180] loss: 0.265 [23, 240] loss: 0.287 [23, 300] loss: 0.267 [23, 360] loss: 0.303 Epoch: 23 -> Loss: 0.45448166132 Epoch: 23 -> Test Accuracy: 85.46 [24, 60] loss: 0.241 [24, 120] loss: 0.261 [24, 180] loss: 0.277 [24, 240] loss: 0.274 [24, 300] loss: 0.285 [24, 360] loss: 0.289 Epoch: 24 -> Loss: 0.220754593611 Epoch: 24 -> Test Accuracy: 85.4 [25, 60] loss: 0.256 [25, 120] loss: 0.249 [25, 180] loss: 0.258 [25, 240] loss: 0.273 [25, 300] loss: 0.278 [25, 360] loss: 0.304 Epoch: 25 -> Loss: 0.273232907057 Epoch: 25 -> Test Accuracy: 85.32 [26, 60] loss: 0.250 [26, 120] loss: 0.254 [26, 180] loss: 0.270 [26, 240] loss: 0.270 [26, 300] loss: 0.286 [26, 360] loss: 0.270 Epoch: 26 -> Loss: 0.186416223645 Epoch: 26 -> Test Accuracy: 86.0 [27, 60] loss: 0.246 [27, 120] loss: 0.254 [27, 180] loss: 0.262 [27, 240] loss: 0.277 [27, 300] loss: 0.294 [27, 360] loss: 0.298 Epoch: 27 -> Loss: 0.315597355366 Epoch: 27 -> Test Accuracy: 85.52 [28, 60] loss: 0.252 [28, 120] loss: 0.255 [28, 180] loss: 0.255 [28, 240] loss: 0.277 [28, 300] loss: 0.280 [28, 360] loss: 0.274 Epoch: 28 -> Loss: 0.28716173768 Epoch: 28 -> Test Accuracy: 86.13 [29, 60] loss: 0.251 [29, 120] loss: 0.249 [29, 180] loss: 0.275 [29, 240] loss: 0.275 [29, 300] loss: 0.293 [29, 360] loss: 0.266 Epoch: 29 -> Loss: 0.236902907491 Epoch: 29 -> Test Accuracy: 86.14 [30, 60] loss: 0.242 [30, 120] loss: 0.254 [30, 180] loss: 0.268 [30, 240] loss: 0.270 [30, 300] loss: 0.285 [30, 360] loss: 0.262 Epoch: 30 -> Loss: 0.310633897781 Epoch: 30 -> Test Accuracy: 85.46 [31, 60] loss: 0.252 [31, 120] loss: 0.249 [31, 180] loss: 0.265 [31, 240] loss: 0.274 [31, 300] loss: 0.278 [31, 360] loss: 0.285 Epoch: 31 -> Loss: 0.324928581715 Epoch: 31 -> Test Accuracy: 86.2 [32, 60] loss: 0.248 [32, 120] loss: 0.260 [32, 180] loss: 0.277 [32, 240] loss: 0.266 [32, 300] loss: 0.275 [32, 360] loss: 0.281 Epoch: 32 -> Loss: 0.217931956053 Epoch: 32 -> Test Accuracy: 85.69 [33, 60] loss: 0.234 [33, 120] loss: 0.257 [33, 180] loss: 0.261 [33, 240] loss: 0.263 [33, 300] loss: 0.284 [33, 360] loss: 0.290 Epoch: 33 -> Loss: 0.226124957204 Epoch: 33 -> Test Accuracy: 85.89 [34, 60] loss: 0.234 [34, 120] loss: 0.247 [34, 180] loss: 0.266 [34, 240] loss: 0.272 [34, 300] loss: 0.284 [34, 360] loss: 0.280 Epoch: 34 -> Loss: 0.312292128801 Epoch: 34 -> Test Accuracy: 85.61 [35, 60] loss: 0.239 [35, 120] loss: 0.258 [35, 180] loss: 0.267 [35, 240] loss: 0.266 [35, 300] loss: 0.276 [35, 360] loss: 0.288 Epoch: 35 -> Loss: 0.232284829021 Epoch: 35 -> Test Accuracy: 86.26 [36, 60] loss: 0.203 [36, 120] loss: 0.188 [36, 180] loss: 0.181 [36, 240] loss: 0.160 [36, 300] loss: 0.169 [36, 360] loss: 0.176 Epoch: 36 -> Loss: 0.147477537394 Epoch: 36 -> Test Accuracy: 88.2 [37, 60] loss: 0.141 [37, 120] loss: 0.147 [37, 180] loss: 0.145 [37, 240] loss: 0.147 [37, 300] loss: 0.150 [37, 360] loss: 0.148 Epoch: 37 -> Loss: 0.127799779177 Epoch: 37 -> Test Accuracy: 88.26 [38, 60] loss: 0.130 [38, 120] loss: 0.131 [38, 180] loss: 0.138 [38, 240] loss: 0.145 [38, 300] loss: 0.137 [38, 360] loss: 0.143 Epoch: 38 -> Loss: 0.101877167821 Epoch: 38 -> Test Accuracy: 88.04 [39, 60] loss: 0.117 [39, 120] loss: 0.120 [39, 180] loss: 0.125 [39, 240] loss: 0.116 [39, 300] loss: 0.135 [39, 360] loss: 0.136 Epoch: 39 -> Loss: 0.118559643626 Epoch: 39 -> Test Accuracy: 88.24 [40, 60] loss: 0.113 [40, 120] loss: 0.117 [40, 180] loss: 0.118 [40, 240] loss: 0.120 [40, 300] loss: 0.124 [40, 360] loss: 0.123 Epoch: 40 -> Loss: 0.0976566374302 Epoch: 40 -> Test Accuracy: 88.12 [41, 60] loss: 0.113 [41, 120] loss: 0.112 [41, 180] loss: 0.115 [41, 240] loss: 0.108 [41, 300] loss: 0.119 [41, 360] loss: 0.122 Epoch: 41 -> Loss: 0.0460130013525 Epoch: 41 -> Test Accuracy: 87.94 [42, 60] loss: 0.100 [42, 120] loss: 0.103 [42, 180] loss: 0.112 [42, 240] loss: 0.111 [42, 300] loss: 0.116 [42, 360] loss: 0.119 Epoch: 42 -> Loss: 0.0915532708168 Epoch: 42 -> Test Accuracy: 88.0 [43, 60] loss: 0.093 [43, 120] loss: 0.104 [43, 180] loss: 0.107 [43, 240] loss: 0.107 [43, 300] loss: 0.120 [43, 360] loss: 0.121 Epoch: 43 -> Loss: 0.0771202594042 Epoch: 43 -> Test Accuracy: 87.67 [44, 60] loss: 0.103 [44, 120] loss: 0.101 [44, 180] loss: 0.104 [44, 240] loss: 0.109 [44, 300] loss: 0.111 [44, 360] loss: 0.122 Epoch: 44 -> Loss: 0.1015945822 Epoch: 44 -> Test Accuracy: 88.09 [45, 60] loss: 0.098 [45, 120] loss: 0.104 [45, 180] loss: 0.102 [45, 240] loss: 0.109 [45, 300] loss: 0.119 [45, 360] loss: 0.109 Epoch: 45 -> Loss: 0.143879905343 Epoch: 45 -> Test Accuracy: 88.02 [46, 60] loss: 0.092 [46, 120] loss: 0.103 [46, 180] loss: 0.105 [46, 240] loss: 0.113 [46, 300] loss: 0.104 [46, 360] loss: 0.113 Epoch: 46 -> Loss: 0.152754217386 Epoch: 46 -> Test Accuracy: 87.93 [47, 60] loss: 0.100 [47, 120] loss: 0.107 [47, 180] loss: 0.110 [47, 240] loss: 0.109 [47, 300] loss: 0.113 [47, 360] loss: 0.117 Epoch: 47 -> Loss: 0.180139839649 Epoch: 47 -> Test Accuracy: 87.57 [48, 60] loss: 0.093 [48, 120] loss: 0.100 [48, 180] loss: 0.095 [48, 240] loss: 0.121 [48, 300] loss: 0.111 [48, 360] loss: 0.114 Epoch: 48 -> Loss: 0.161243930459 Epoch: 48 -> Test Accuracy: 87.3 [49, 60] loss: 0.095 [49, 120] loss: 0.099 [49, 180] loss: 0.109 [49, 240] loss: 0.107 [49, 300] loss: 0.120 [49, 360] loss: 0.119 Epoch: 49 -> Loss: 0.136429190636 Epoch: 49 -> Test Accuracy: 87.67 [50, 60] loss: 0.094 [50, 120] loss: 0.104 [50, 180] loss: 0.102 [50, 240] loss: 0.105 [50, 300] loss: 0.113 [50, 360] loss: 0.120 Epoch: 50 -> Loss: 0.105074964464 Epoch: 50 -> Test Accuracy: 87.79 [51, 60] loss: 0.102 [51, 120] loss: 0.106 [51, 180] loss: 0.109 [51, 240] loss: 0.103 [51, 300] loss: 0.108 [51, 360] loss: 0.122 Epoch: 51 -> Loss: 0.130261033773 Epoch: 51 -> Test Accuracy: 87.58 [52, 60] loss: 0.099 [52, 120] loss: 0.096 [52, 180] loss: 0.109 [52, 240] loss: 0.105 [52, 300] loss: 0.123 [52, 360] loss: 0.120 Epoch: 52 -> Loss: 0.105667151511 Epoch: 52 -> Test Accuracy: 87.7 [53, 60] loss: 0.100 [53, 120] loss: 0.106 [53, 180] loss: 0.101 [53, 240] loss: 0.113 [53, 300] loss: 0.116 [53, 360] loss: 0.127 Epoch: 53 -> Loss: 0.101517722011 Epoch: 53 -> Test Accuracy: 87.69 [54, 60] loss: 0.103 [54, 120] loss: 0.101 [54, 180] loss: 0.111 [54, 240] loss: 0.111 [54, 300] loss: 0.115 [54, 360] loss: 0.109 Epoch: 54 -> Loss: 0.0614903457463 Epoch: 54 -> Test Accuracy: 87.23 [55, 60] loss: 0.097 [55, 120] loss: 0.096 [55, 180] loss: 0.104 [55, 240] loss: 0.109 [55, 300] loss: 0.121 [55, 360] loss: 0.121 Epoch: 55 -> Loss: 0.0488305650651 Epoch: 55 -> Test Accuracy: 87.71 [56, 60] loss: 0.106 [56, 120] loss: 0.105 [56, 180] loss: 0.109 [56, 240] loss: 0.105 [56, 300] loss: 0.113 [56, 360] loss: 0.121 Epoch: 56 -> Loss: 0.179651007056 Epoch: 56 -> Test Accuracy: 87.39 [57, 60] loss: 0.105 [57, 120] loss: 0.108 [57, 180] loss: 0.112 [57, 240] loss: 0.114 [57, 300] loss: 0.104 [57, 360] loss: 0.107 Epoch: 57 -> Loss: 0.133594423532 Epoch: 57 -> Test Accuracy: 87.06 [58, 60] loss: 0.105 [58, 120] loss: 0.103 [58, 180] loss: 0.106 [58, 240] loss: 0.104 [58, 300] loss: 0.116 [58, 360] loss: 0.119 Epoch: 58 -> Loss: 0.113112285733 Epoch: 58 -> Test Accuracy: 87.46 [59, 60] loss: 0.100 [59, 120] loss: 0.105 [59, 180] loss: 0.112 [59, 240] loss: 0.110 [59, 300] loss: 0.108 [59, 360] loss: 0.123 Epoch: 59 -> Loss: 0.120920315385 Epoch: 59 -> Test Accuracy: 87.41 [60, 60] loss: 0.102 [60, 120] loss: 0.108 [60, 180] loss: 0.110 [60, 240] loss: 0.107 [60, 300] loss: 0.115 [60, 360] loss: 0.113 Epoch: 60 -> Loss: 0.128071188927 Epoch: 60 -> Test Accuracy: 86.91 [61, 60] loss: 0.109 [61, 120] loss: 0.100 [61, 180] loss: 0.107 [61, 240] loss: 0.116 [61, 300] loss: 0.110 [61, 360] loss: 0.108 Epoch: 61 -> Loss: 0.24754679203 Epoch: 61 -> Test Accuracy: 87.13 [62, 60] loss: 0.099 [62, 120] loss: 0.099 [62, 180] loss: 0.098 [62, 240] loss: 0.107 [62, 300] loss: 0.117 [62, 360] loss: 0.123 Epoch: 62 -> Loss: 0.140629321337 Epoch: 62 -> Test Accuracy: 87.19 [63, 60] loss: 0.099 [63, 120] loss: 0.109 [63, 180] loss: 0.096 [63, 240] loss: 0.106 [63, 300] loss: 0.112 [63, 360] loss: 0.113 Epoch: 63 -> Loss: 0.116040453315 Epoch: 63 -> Test Accuracy: 87.03 [64, 60] loss: 0.099 [64, 120] loss: 0.091 [64, 180] loss: 0.098 [64, 240] loss: 0.107 [64, 300] loss: 0.116 [64, 360] loss: 0.118 Epoch: 64 -> Loss: 0.0881711989641 Epoch: 64 -> Test Accuracy: 87.67 [65, 60] loss: 0.094 [65, 120] loss: 0.101 [65, 180] loss: 0.103 [65, 240] loss: 0.104 [65, 300] loss: 0.101 [65, 360] loss: 0.114 Epoch: 65 -> Loss: 0.0817269459367 Epoch: 65 -> Test Accuracy: 87.36 [66, 60] loss: 0.097 [66, 120] loss: 0.106 [66, 180] loss: 0.102 [66, 240] loss: 0.109 [66, 300] loss: 0.111 [66, 360] loss: 0.114 Epoch: 66 -> Loss: 0.164013296366 Epoch: 66 -> Test Accuracy: 87.58 [67, 60] loss: 0.103 [67, 120] loss: 0.105 [67, 180] loss: 0.103 [67, 240] loss: 0.114 [67, 300] loss: 0.108 [67, 360] loss: 0.114 Epoch: 67 -> Loss: 0.0772556811571 Epoch: 67 -> Test Accuracy: 86.82 [68, 60] loss: 0.100 [68, 120] loss: 0.103 [68, 180] loss: 0.108 [68, 240] loss: 0.100 [68, 300] loss: 0.110 [68, 360] loss: 0.113 Epoch: 68 -> Loss: 0.0822162479162 Epoch: 68 -> Test Accuracy: 87.56 [69, 60] loss: 0.097 [69, 120] loss: 0.098 [69, 180] loss: 0.101 [69, 240] loss: 0.104 [69, 300] loss: 0.116 [69, 360] loss: 0.111 Epoch: 69 -> Loss: 0.102196291089 Epoch: 69 -> Test Accuracy: 87.01 [70, 60] loss: 0.100 [70, 120] loss: 0.093 [70, 180] loss: 0.108 [70, 240] loss: 0.109 [70, 300] loss: 0.104 [70, 360] loss: 0.114 Epoch: 70 -> Loss: 0.189903616905 Epoch: 70 -> Test Accuracy: 87.45 [71, 60] loss: 0.074 [71, 120] loss: 0.071 [71, 180] loss: 0.068 [71, 240] loss: 0.064 [71, 300] loss: 0.062 [71, 360] loss: 0.062 Epoch: 71 -> Loss: 0.0460370704532 Epoch: 71 -> Test Accuracy: 88.58 [72, 60] loss: 0.054 [72, 120] loss: 0.056 [72, 180] loss: 0.053 [72, 240] loss: 0.050 [72, 300] loss: 0.060 [72, 360] loss: 0.056 Epoch: 72 -> Loss: 0.0404352359474 Epoch: 72 -> Test Accuracy: 88.54 [73, 60] loss: 0.049 [73, 120] loss: 0.044 [73, 180] loss: 0.052 [73, 240] loss: 0.051 [73, 300] loss: 0.051 [73, 360] loss: 0.048 Epoch: 73 -> Loss: 0.0345988273621 Epoch: 73 -> Test Accuracy: 88.76 [74, 60] loss: 0.042 [74, 120] loss: 0.047 [74, 180] loss: 0.045 [74, 240] loss: 0.048 [74, 300] loss: 0.047 [74, 360] loss: 0.042 Epoch: 74 -> Loss: 0.020695855841 Epoch: 74 -> Test Accuracy: 88.74 [75, 60] loss: 0.043 [75, 120] loss: 0.044 [75, 180] loss: 0.041 [75, 240] loss: 0.044 [75, 300] loss: 0.044 [75, 360] loss: 0.045 Epoch: 75 -> Loss: 0.0674355328083 Epoch: 75 -> Test Accuracy: 88.52 [76, 60] loss: 0.043 [76, 120] loss: 0.041 [76, 180] loss: 0.038 [76, 240] loss: 0.039 [76, 300] loss: 0.042 [76, 360] loss: 0.043 Epoch: 76 -> Loss: 0.0397047698498 Epoch: 76 -> Test Accuracy: 88.69 [77, 60] loss: 0.041 [77, 120] loss: 0.041 [77, 180] loss: 0.037 [77, 240] loss: 0.041 [77, 300] loss: 0.044 [77, 360] loss: 0.040 Epoch: 77 -> Loss: 0.0739384442568 Epoch: 77 -> Test Accuracy: 88.88 [78, 60] loss: 0.036 [78, 120] loss: 0.038 [78, 180] loss: 0.041 [78, 240] loss: 0.040 [78, 300] loss: 0.039 [78, 360] loss: 0.038 Epoch: 78 -> Loss: 0.0402158088982 Epoch: 78 -> Test Accuracy: 88.58 [79, 60] loss: 0.037 [79, 120] loss: 0.041 [79, 180] loss: 0.035 [79, 240] loss: 0.037 [79, 300] loss: 0.037 [79, 360] loss: 0.037 Epoch: 79 -> Loss: 0.0498256273568 Epoch: 79 -> Test Accuracy: 88.7 [80, 60] loss: 0.037 [80, 120] loss: 0.035 [80, 180] loss: 0.037 [80, 240] loss: 0.035 [80, 300] loss: 0.040 [80, 360] loss: 0.037 Epoch: 80 -> Loss: 0.0296311341226 Epoch: 80 -> Test Accuracy: 88.63 [81, 60] loss: 0.035 [81, 120] loss: 0.036 [81, 180] loss: 0.039 [81, 240] loss: 0.037 [81, 300] loss: 0.034 [81, 360] loss: 0.035 Epoch: 81 -> Loss: 0.050663150847 Epoch: 81 -> Test Accuracy: 88.52 [82, 60] loss: 0.033 [82, 120] loss: 0.031 [82, 180] loss: 0.035 [82, 240] loss: 0.035 [82, 300] loss: 0.038 [82, 360] loss: 0.035 Epoch: 82 -> Loss: 0.021719366312 Epoch: 82 -> Test Accuracy: 88.72 [83, 60] loss: 0.032 [83, 120] loss: 0.035 [83, 180] loss: 0.037 [83, 240] loss: 0.033 [83, 300] loss: 0.034 [83, 360] loss: 0.034 Epoch: 83 -> Loss: 0.021537065506 Epoch: 83 -> Test Accuracy: 88.86 [84, 60] loss: 0.031 [84, 120] loss: 0.034 [84, 180] loss: 0.036 [84, 240] loss: 0.036 [84, 300] loss: 0.034 [84, 360] loss: 0.032 Epoch: 84 -> Loss: 0.0686306804419 Epoch: 84 -> Test Accuracy: 88.78 [85, 60] loss: 0.032 [85, 120] loss: 0.032 [85, 180] loss: 0.032 [85, 240] loss: 0.032 [85, 300] loss: 0.032 [85, 360] loss: 0.032 Epoch: 85 -> Loss: 0.038092110306 Epoch: 85 -> Test Accuracy: 88.81 [86, 60] loss: 0.029 [86, 120] loss: 0.030 [86, 180] loss: 0.029 [86, 240] loss: 0.029 [86, 300] loss: 0.029 [86, 360] loss: 0.029 Epoch: 86 -> Loss: 0.0370008982718 Epoch: 86 -> Test Accuracy: 88.88 [87, 60] loss: 0.029 [87, 120] loss: 0.027 [87, 180] loss: 0.028 [87, 240] loss: 0.027 [87, 300] loss: 0.031 [87, 360] loss: 0.030 Epoch: 87 -> Loss: 0.0638825148344 Epoch: 87 -> Test Accuracy: 88.99 [88, 60] loss: 0.030 [88, 120] loss: 0.028 [88, 180] loss: 0.027 [88, 240] loss: 0.028 [88, 300] loss: 0.025 [88, 360] loss: 0.028 Epoch: 88 -> Loss: 0.0306285060942 Epoch: 88 -> Test Accuracy: 88.68 [89, 60] loss: 0.027 [89, 120] loss: 0.027 [89, 180] loss: 0.029 [89, 240] loss: 0.030 [89, 300] loss: 0.028 [89, 360] loss: 0.030 Epoch: 89 -> Loss: 0.0529460385442 Epoch: 89 -> Test Accuracy: 88.88 [90, 60] loss: 0.027 [90, 120] loss: 0.028 [90, 180] loss: 0.029 [90, 240] loss: 0.029 [90, 300] loss: 0.029 [90, 360] loss: 0.025 Epoch: 90 -> Loss: 0.0197838954628 Epoch: 90 -> Test Accuracy: 88.89 [91, 60] loss: 0.026 [91, 120] loss: 0.029 [91, 180] loss: 0.028 [91, 240] loss: 0.028 [91, 300] loss: 0.026 [91, 360] loss: 0.026 Epoch: 91 -> Loss: 0.024331022054 Epoch: 91 -> Test Accuracy: 88.9 [92, 60] loss: 0.027 [92, 120] loss: 0.028 [92, 180] loss: 0.026 [92, 240] loss: 0.026 [92, 300] loss: 0.025 [92, 360] loss: 0.029 Epoch: 92 -> Loss: 0.0443188846111 Epoch: 92 -> Test Accuracy: 88.86 [93, 60] loss: 0.025 [93, 120] loss: 0.027 [93, 180] loss: 0.028 [93, 240] loss: 0.027 [93, 300] loss: 0.029 [93, 360] loss: 0.029 Epoch: 93 -> Loss: 0.0265762563795 Epoch: 93 -> Test Accuracy: 88.8 [94, 60] loss: 0.026 [94, 120] loss: 0.027 [94, 180] loss: 0.027 [94, 240] loss: 0.026 [94, 300] loss: 0.027 [94, 360] loss: 0.026 Epoch: 94 -> Loss: 0.0204644501209 Epoch: 94 -> Test Accuracy: 88.72 [95, 60] loss: 0.025 [95, 120] loss: 0.027 [95, 180] loss: 0.028 [95, 240] loss: 0.028 [95, 300] loss: 0.027 [95, 360] loss: 0.027 Epoch: 95 -> Loss: 0.0251078605652 Epoch: 95 -> Test Accuracy: 88.82 [96, 60] loss: 0.025 [96, 120] loss: 0.025 [96, 180] loss: 0.028 [96, 240] loss: 0.025 [96, 300] loss: 0.027 [96, 360] loss: 0.027 Epoch: 96 -> Loss: 0.0565241165459 Epoch: 96 -> Test Accuracy: 88.9 [97, 60] loss: 0.029 [97, 120] loss: 0.026 [97, 180] loss: 0.025 [97, 240] loss: 0.027 [97, 300] loss: 0.027 [97, 360] loss: 0.026 Epoch: 97 -> Loss: 0.0237293429673 Epoch: 97 -> Test Accuracy: 88.82 [98, 60] loss: 0.023 [98, 120] loss: 0.025 [98, 180] loss: 0.027 [98, 240] loss: 0.026 [98, 300] loss: 0.025 [98, 360] loss: 0.027 Epoch: 98 -> Loss: 0.0092001138255 Epoch: 98 -> Test Accuracy: 88.76 [99, 60] loss: 0.024 [99, 120] loss: 0.029 [99, 180] loss: 0.025 [99, 240] loss: 0.025 [99, 300] loss: 0.026 [99, 360] loss: 0.027 Epoch: 99 -> Loss: 0.0157920122147 Epoch: 99 -> Test Accuracy: 88.78 [100, 60] loss: 0.025 [100, 120] loss: 0.025 [100, 180] loss: 0.025 [100, 240] loss: 0.025 [100, 300] loss: 0.026 [100, 360] loss: 0.024 Epoch: 100 -> Loss: 0.0201563090086 Epoch: 100 -> Test Accuracy: 88.82 Finished Training [1, 60] loss: 1.863 [1, 120] loss: 1.657 [1, 180] loss: 1.573 [1, 240] loss: 1.547 [1, 300] loss: 1.518 [1, 360] loss: 1.465 Epoch: 1 -> Loss: 1.38100779057 Epoch: 1 -> Test Accuracy: 42.51 [2, 60] loss: 1.455 [2, 120] loss: 1.455 [2, 180] loss: 1.424 [2, 240] loss: 1.439 [2, 300] loss: 1.431 [2, 360] loss: 1.394 Epoch: 2 -> Loss: 1.57546544075 Epoch: 2 -> Test Accuracy: 46.54 [3, 60] loss: 1.372 [3, 120] loss: 1.386 [3, 180] loss: 1.390 [3, 240] loss: 1.368 [3, 300] loss: 1.359 [3, 360] loss: 1.359 Epoch: 3 -> Loss: 1.44619822502 Epoch: 3 -> Test Accuracy: 46.99 [4, 60] loss: 1.343 [4, 120] loss: 1.357 [4, 180] loss: 1.338 [4, 240] loss: 1.347 [4, 300] loss: 1.311 [4, 360] loss: 1.337 Epoch: 4 -> Loss: 1.35329127312 Epoch: 4 -> Test Accuracy: 48.98 [5, 60] loss: 1.329 [5, 120] loss: 1.306 [5, 180] loss: 1.312 [5, 240] loss: 1.317 [5, 300] loss: 1.333 [5, 360] loss: 1.317 Epoch: 5 -> Loss: 1.33428633213 Epoch: 5 -> Test Accuracy: 48.82 [6, 60] loss: 1.311 [6, 120] loss: 1.305 [6, 180] loss: 1.277 [6, 240] loss: 1.309 [6, 300] loss: 1.308 [6, 360] loss: 1.283 Epoch: 6 -> Loss: 1.38868832588 Epoch: 6 -> Test Accuracy: 48.68 [7, 60] loss: 1.287 [7, 120] loss: 1.294 [7, 180] loss: 1.272 [7, 240] loss: 1.288 [7, 300] loss: 1.297 [7, 360] loss: 1.285 Epoch: 7 -> Loss: 1.12669825554 Epoch: 7 -> Test Accuracy: 49.95 [8, 60] loss: 1.289 [8, 120] loss: 1.275 [8, 180] loss: 1.278 [8, 240] loss: 1.263 [8, 300] loss: 1.283 [8, 360] loss: 1.283 Epoch: 8 -> Loss: 1.37220048904 Epoch: 8 -> Test Accuracy: 50.36 [9, 60] loss: 1.268 [9, 120] loss: 1.257 [9, 180] loss: 1.257 [9, 240] loss: 1.270 [9, 300] loss: 1.275 [9, 360] loss: 1.271 Epoch: 9 -> Loss: 1.19314074516 Epoch: 9 -> Test Accuracy: 49.89 [10, 60] loss: 1.278 [10, 120] loss: 1.246 [10, 180] loss: 1.270 [10, 240] loss: 1.252 [10, 300] loss: 1.265 [10, 360] loss: 1.269 Epoch: 10 -> Loss: 1.18051970005 Epoch: 10 -> Test Accuracy: 50.67 [11, 60] loss: 1.248 [11, 120] loss: 1.282 [11, 180] loss: 1.244 [11, 240] loss: 1.267 [11, 300] loss: 1.260 [11, 360] loss: 1.241 Epoch: 11 -> Loss: 1.45489859581 Epoch: 11 -> Test Accuracy: 50.52 [12, 60] loss: 1.276 [12, 120] loss: 1.252 [12, 180] loss: 1.259 [12, 240] loss: 1.249 [12, 300] loss: 1.243 [12, 360] loss: 1.251 Epoch: 12 -> Loss: 1.29271221161 Epoch: 12 -> Test Accuracy: 49.86 [13, 60] loss: 1.252 [13, 120] loss: 1.256 [13, 180] loss: 1.255 [13, 240] loss: 1.249 [13, 300] loss: 1.249 [13, 360] loss: 1.262 Epoch: 13 -> Loss: 1.41856431961 Epoch: 13 -> Test Accuracy: 50.32 [14, 60] loss: 1.262 [14, 120] loss: 1.239 [14, 180] loss: 1.260 [14, 240] loss: 1.256 [14, 300] loss: 1.254 [14, 360] loss: 1.233 Epoch: 14 -> Loss: 1.20106911659 Epoch: 14 -> Test Accuracy: 49.65 [15, 60] loss: 1.250 [15, 120] loss: 1.250 [15, 180] loss: 1.234 [15, 240] loss: 1.224 [15, 300] loss: 1.250 [15, 360] loss: 1.235 Epoch: 15 -> Loss: 1.21732747555 Epoch: 15 -> Test Accuracy: 51.29 [16, 60] loss: 1.247 [16, 120] loss: 1.246 [16, 180] loss: 1.217 [16, 240] loss: 1.224 [16, 300] loss: 1.253 [16, 360] loss: 1.238 Epoch: 16 -> Loss: 1.28469765186 Epoch: 16 -> Test Accuracy: 51.23 [17, 60] loss: 1.226 [17, 120] loss: 1.239 [17, 180] loss: 1.248 [17, 240] loss: 1.240 [17, 300] loss: 1.244 [17, 360] loss: 1.251 Epoch: 17 -> Loss: 1.24963212013 Epoch: 17 -> Test Accuracy: 51.43 [18, 60] loss: 1.245 [18, 120] loss: 1.241 [18, 180] loss: 1.260 [18, 240] loss: 1.241 [18, 300] loss: 1.249 [18, 360] loss: 1.231 Epoch: 18 -> Loss: 1.35986924171 Epoch: 18 -> Test Accuracy: 50.88 [19, 60] loss: 1.251 [19, 120] loss: 1.233 [19, 180] loss: 1.229 [19, 240] loss: 1.237 [19, 300] loss: 1.233 [19, 360] loss: 1.231 Epoch: 19 -> Loss: 1.27556943893 Epoch: 19 -> Test Accuracy: 50.71 [20, 60] loss: 1.236 [20, 120] loss: 1.227 [20, 180] loss: 1.237 [20, 240] loss: 1.235 [20, 300] loss: 1.200 [20, 360] loss: 1.246 Epoch: 20 -> Loss: 1.04955816269 Epoch: 20 -> Test Accuracy: 50.77 [21, 60] loss: 1.234 [21, 120] loss: 1.219 [21, 180] loss: 1.241 [21, 240] loss: 1.230 [21, 300] loss: 1.253 [21, 360] loss: 1.221 Epoch: 21 -> Loss: 1.27590477467 Epoch: 21 -> Test Accuracy: 51.65 [22, 60] loss: 1.239 [22, 120] loss: 1.219 [22, 180] loss: 1.216 [22, 240] loss: 1.223 [22, 300] loss: 1.239 [22, 360] loss: 1.237 Epoch: 22 -> Loss: 1.14897048473 Epoch: 22 -> Test Accuracy: 52.7 [23, 60] loss: 1.232 [23, 120] loss: 1.231 [23, 180] loss: 1.215 [23, 240] loss: 1.230 [23, 300] loss: 1.220 [23, 360] loss: 1.253 Epoch: 23 -> Loss: 1.38017296791 Epoch: 23 -> Test Accuracy: 51.58 [24, 60] loss: 1.223 [24, 120] loss: 1.242 [24, 180] loss: 1.219 [24, 240] loss: 1.231 [24, 300] loss: 1.230 [24, 360] loss: 1.232 Epoch: 24 -> Loss: 1.33723008633 Epoch: 24 -> Test Accuracy: 50.85 [25, 60] loss: 1.218 [25, 120] loss: 1.220 [25, 180] loss: 1.244 [25, 240] loss: 1.210 [25, 300] loss: 1.245 [25, 360] loss: 1.240 Epoch: 25 -> Loss: 1.19878721237 Epoch: 25 -> Test Accuracy: 50.58 [26, 60] loss: 1.242 [26, 120] loss: 1.240 [26, 180] loss: 1.208 [26, 240] loss: 1.222 [26, 300] loss: 1.225 [26, 360] loss: 1.220 Epoch: 26 -> Loss: 1.30900859833 Epoch: 26 -> Test Accuracy: 52.41 [27, 60] loss: 1.221 [27, 120] loss: 1.225 [27, 180] loss: 1.224 [27, 240] loss: 1.233 [27, 300] loss: 1.218 [27, 360] loss: 1.201 Epoch: 27 -> Loss: 1.29304289818 Epoch: 27 -> Test Accuracy: 51.48 [28, 60] loss: 1.235 [28, 120] loss: 1.200 [28, 180] loss: 1.212 [28, 240] loss: 1.229 [28, 300] loss: 1.252 [28, 360] loss: 1.250 Epoch: 28 -> Loss: 1.21757674217 Epoch: 28 -> Test Accuracy: 50.61 [29, 60] loss: 1.236 [29, 120] loss: 1.219 [29, 180] loss: 1.221 [29, 240] loss: 1.235 [29, 300] loss: 1.239 [29, 360] loss: 1.212 Epoch: 29 -> Loss: 1.29875969887 Epoch: 29 -> Test Accuracy: 51.41 [30, 60] loss: 1.200 [30, 120] loss: 1.228 [30, 180] loss: 1.215 [30, 240] loss: 1.227 [30, 300] loss: 1.237 [30, 360] loss: 1.239 Epoch: 30 -> Loss: 1.42523491383 Epoch: 30 -> Test Accuracy: 52.18 [31, 60] loss: 1.216 [31, 120] loss: 1.223 [31, 180] loss: 1.213 [31, 240] loss: 1.239 [31, 300] loss: 1.220 [31, 360] loss: 1.222 Epoch: 31 -> Loss: 1.45415902138 Epoch: 31 -> Test Accuracy: 51.43 [32, 60] loss: 1.229 [32, 120] loss: 1.199 [32, 180] loss: 1.220 [32, 240] loss: 1.234 [32, 300] loss: 1.241 [32, 360] loss: 1.227 Epoch: 32 -> Loss: 1.24754357338 Epoch: 32 -> Test Accuracy: 50.31 [33, 60] loss: 1.212 [33, 120] loss: 1.217 [33, 180] loss: 1.222 [33, 240] loss: 1.230 [33, 300] loss: 1.222 [33, 360] loss: 1.223 Epoch: 33 -> Loss: 1.21123230457 Epoch: 33 -> Test Accuracy: 52.15 [34, 60] loss: 1.231 [34, 120] loss: 1.237 [34, 180] loss: 1.204 [34, 240] loss: 1.231 [34, 300] loss: 1.209 [34, 360] loss: 1.218 Epoch: 34 -> Loss: 1.27467799187 Epoch: 34 -> Test Accuracy: 51.37 [35, 60] loss: 1.201 [35, 120] loss: 1.208 [35, 180] loss: 1.224 [35, 240] loss: 1.230 [35, 300] loss: 1.244 [35, 360] loss: 1.226 Epoch: 35 -> Loss: 1.30367970467 Epoch: 35 -> Test Accuracy: 51.49 [36, 60] loss: 1.162 [36, 120] loss: 1.125 [36, 180] loss: 1.105 [36, 240] loss: 1.097 [36, 300] loss: 1.100 [36, 360] loss: 1.081 Epoch: 36 -> Loss: 1.13254475594 Epoch: 36 -> Test Accuracy: 56.02 [37, 60] loss: 1.088 [37, 120] loss: 1.099 [37, 180] loss: 1.086 [37, 240] loss: 1.068 [37, 300] loss: 1.099 [37, 360] loss: 1.099 Epoch: 37 -> Loss: 1.07157111168 Epoch: 37 -> Test Accuracy: 56.53 [38, 60] loss: 1.071 [38, 120] loss: 1.076 [38, 180] loss: 1.077 [38, 240] loss: 1.080 [38, 300] loss: 1.099 [38, 360] loss: 1.079 Epoch: 38 -> Loss: 1.01064562798 Epoch: 38 -> Test Accuracy: 56.76 [39, 60] loss: 1.077 [39, 120] loss: 1.070 [39, 180] loss: 1.099 [39, 240] loss: 1.090 [39, 300] loss: 1.065 [39, 360] loss: 1.082 Epoch: 39 -> Loss: 0.90257537365 Epoch: 39 -> Test Accuracy: 56.6 [40, 60] loss: 1.070 [40, 120] loss: 1.087 [40, 180] loss: 1.071 [40, 240] loss: 1.080 [40, 300] loss: 1.086 [40, 360] loss: 1.083 Epoch: 40 -> Loss: 0.987020790577 Epoch: 40 -> Test Accuracy: 56.6 [41, 60] loss: 1.051 [41, 120] loss: 1.089 [41, 180] loss: 1.086 [41, 240] loss: 1.068 [41, 300] loss: 1.077 [41, 360] loss: 1.081 Epoch: 41 -> Loss: 1.08318543434 Epoch: 41 -> Test Accuracy: 57.24 [42, 60] loss: 1.083 [42, 120] loss: 1.054 [42, 180] loss: 1.072 [42, 240] loss: 1.069 [42, 300] loss: 1.075 [42, 360] loss: 1.086 Epoch: 42 -> Loss: 0.9835947752 Epoch: 42 -> Test Accuracy: 56.93 [43, 60] loss: 1.056 [43, 120] loss: 1.080 [43, 180] loss: 1.067 [43, 240] loss: 1.095 [43, 300] loss: 1.076 [43, 360] loss: 1.069 Epoch: 43 -> Loss: 1.08890414238 Epoch: 43 -> Test Accuracy: 57.08 [44, 60] loss: 1.078 [44, 120] loss: 1.065 [44, 180] loss: 1.072 [44, 240] loss: 1.089 [44, 300] loss: 1.072 [44, 360] loss: 1.075 Epoch: 44 -> Loss: 1.07605016232 Epoch: 44 -> Test Accuracy: 57.03 [45, 60] loss: 1.064 [45, 120] loss: 1.074 [45, 180] loss: 1.053 [45, 240] loss: 1.078 [45, 300] loss: 1.051 [45, 360] loss: 1.095 Epoch: 45 -> Loss: 1.16944479942 Epoch: 45 -> Test Accuracy: 57.08 [46, 60] loss: 1.057 [46, 120] loss: 1.080 [46, 180] loss: 1.089 [46, 240] loss: 1.065 [46, 300] loss: 1.076 [46, 360] loss: 1.075 Epoch: 46 -> Loss: 1.18470239639 Epoch: 46 -> Test Accuracy: 56.55 [47, 60] loss: 1.063 [47, 120] loss: 1.064 [47, 180] loss: 1.067 [47, 240] loss: 1.095 [47, 300] loss: 1.055 [47, 360] loss: 1.070 Epoch: 47 -> Loss: 1.24658799171 Epoch: 47 -> Test Accuracy: 57.31 [48, 60] loss: 1.065 [48, 120] loss: 1.077 [48, 180] loss: 1.075 [48, 240] loss: 1.097 [48, 300] loss: 1.085 [48, 360] loss: 1.042 Epoch: 48 -> Loss: 1.10153579712 Epoch: 48 -> Test Accuracy: 57.02 [49, 60] loss: 1.071 [49, 120] loss: 1.100 [49, 180] loss: 1.078 [49, 240] loss: 1.088 [49, 300] loss: 1.048 [49, 360] loss: 1.077 Epoch: 49 -> Loss: 1.06397080421 Epoch: 49 -> Test Accuracy: 55.96 [50, 60] loss: 1.067 [50, 120] loss: 1.063 [50, 180] loss: 1.069 [50, 240] loss: 1.073 [50, 300] loss: 1.082 [50, 360] loss: 1.074 Epoch: 50 -> Loss: 1.01605772972 Epoch: 50 -> Test Accuracy: 56.64 [51, 60] loss: 1.076 [51, 120] loss: 1.099 [51, 180] loss: 1.069 [51, 240] loss: 1.062 [51, 300] loss: 1.081 [51, 360] loss: 1.075 Epoch: 51 -> Loss: 0.962502360344 Epoch: 51 -> Test Accuracy: 56.47 [52, 60] loss: 1.090 [52, 120] loss: 1.079 [52, 180] loss: 1.045 [52, 240] loss: 1.064 [52, 300] loss: 1.083 [52, 360] loss: 1.047 Epoch: 52 -> Loss: 1.1834911108 Epoch: 52 -> Test Accuracy: 56.96 [53, 60] loss: 1.054 [53, 120] loss: 1.051 [53, 180] loss: 1.050 [53, 240] loss: 1.085 [53, 300] loss: 1.069 [53, 360] loss: 1.078 Epoch: 53 -> Loss: 1.04760217667 Epoch: 53 -> Test Accuracy: 57.01 [54, 60] loss: 1.051 [54, 120] loss: 1.065 [54, 180] loss: 1.095 [54, 240] loss: 1.070 [54, 300] loss: 1.064 [54, 360] loss: 1.057 Epoch: 54 -> Loss: 1.19820570946 Epoch: 54 -> Test Accuracy: 57.88 [55, 60] loss: 1.073 [55, 120] loss: 1.072 [55, 180] loss: 1.066 [55, 240] loss: 1.072 [55, 300] loss: 1.062 [55, 360] loss: 1.043 Epoch: 55 -> Loss: 1.05775618553 Epoch: 55 -> Test Accuracy: 56.67 [56, 60] loss: 1.057 [56, 120] loss: 1.084 [56, 180] loss: 1.067 [56, 240] loss: 1.079 [56, 300] loss: 1.062 [56, 360] loss: 1.083 Epoch: 56 -> Loss: 1.13094699383 Epoch: 56 -> Test Accuracy: 57.29 [57, 60] loss: 1.074 [57, 120] loss: 1.066 [57, 180] loss: 1.067 [57, 240] loss: 1.046 [57, 300] loss: 1.060 [57, 360] loss: 1.073 Epoch: 57 -> Loss: 1.11085772514 Epoch: 57 -> Test Accuracy: 56.35 [58, 60] loss: 1.057 [58, 120] loss: 1.077 [58, 180] loss: 1.072 [58, 240] loss: 1.059 [58, 300] loss: 1.078 [58, 360] loss: 1.051 Epoch: 58 -> Loss: 1.01605200768 Epoch: 58 -> Test Accuracy: 56.65 [59, 60] loss: 1.047 [59, 120] loss: 1.072 [59, 180] loss: 1.055 [59, 240] loss: 1.073 [59, 300] loss: 1.057 [59, 360] loss: 1.094 Epoch: 59 -> Loss: 1.15581321716 Epoch: 59 -> Test Accuracy: 57.27 [60, 60] loss: 1.054 [60, 120] loss: 1.061 [60, 180] loss: 1.060 [60, 240] loss: 1.056 [60, 300] loss: 1.073 [60, 360] loss: 1.062 Epoch: 60 -> Loss: 1.01332175732 Epoch: 60 -> Test Accuracy: 56.89 [61, 60] loss: 1.047 [61, 120] loss: 1.066 [61, 180] loss: 1.069 [61, 240] loss: 1.065 [61, 300] loss: 1.076 [61, 360] loss: 1.091 Epoch: 61 -> Loss: 0.988394081593 Epoch: 61 -> Test Accuracy: 56.44 [62, 60] loss: 1.083 [62, 120] loss: 1.070 [62, 180] loss: 1.048 [62, 240] loss: 1.078 [62, 300] loss: 1.059 [62, 360] loss: 1.068 Epoch: 62 -> Loss: 1.20921587944 Epoch: 62 -> Test Accuracy: 57.19 [63, 60] loss: 1.033 [63, 120] loss: 1.043 [63, 180] loss: 1.060 [63, 240] loss: 1.058 [63, 300] loss: 1.086 [63, 360] loss: 1.087 Epoch: 63 -> Loss: 1.16820049286 Epoch: 63 -> Test Accuracy: 57.68 [64, 60] loss: 1.067 [64, 120] loss: 1.063 [64, 180] loss: 1.053 [64, 240] loss: 1.084 [64, 300] loss: 1.055 [64, 360] loss: 1.076 Epoch: 64 -> Loss: 1.00641024113 Epoch: 64 -> Test Accuracy: 56.5 [65, 60] loss: 1.046 [65, 120] loss: 1.086 [65, 180] loss: 1.044 [65, 240] loss: 1.064 [65, 300] loss: 1.075 [65, 360] loss: 1.062 Epoch: 65 -> Loss: 1.07456374168 Epoch: 65 -> Test Accuracy: 57.14 [66, 60] loss: 1.066 [66, 120] loss: 1.067 [66, 180] loss: 1.052 [66, 240] loss: 1.064 [66, 300] loss: 1.057 [66, 360] loss: 1.052 Epoch: 66 -> Loss: 1.03426754475 Epoch: 66 -> Test Accuracy: 56.79 [67, 60] loss: 1.050 [67, 120] loss: 1.056 [67, 180] loss: 1.062 [67, 240] loss: 1.044 [67, 300] loss: 1.083 [67, 360] loss: 1.075 Epoch: 67 -> Loss: 1.13846027851 Epoch: 67 -> Test Accuracy: 57.7 [68, 60] loss: 1.076 [68, 120] loss: 1.053 [68, 180] loss: 1.060 [68, 240] loss: 1.075 [68, 300] loss: 1.052 [68, 360] loss: 1.055 Epoch: 68 -> Loss: 1.14730751514 Epoch: 68 -> Test Accuracy: 56.93 [69, 60] loss: 1.043 [69, 120] loss: 1.051 [69, 180] loss: 1.069 [69, 240] loss: 1.047 [69, 300] loss: 1.064 [69, 360] loss: 1.078 Epoch: 69 -> Loss: 1.09532833099 Epoch: 69 -> Test Accuracy: 57.55 [70, 60] loss: 1.074 [70, 120] loss: 1.060 [70, 180] loss: 1.069 [70, 240] loss: 1.065 [70, 300] loss: 1.055 [70, 360] loss: 1.057 Epoch: 70 -> Loss: 1.0877995491 Epoch: 70 -> Test Accuracy: 57.3 [71, 60] loss: 1.026 [71, 120] loss: 0.990 [71, 180] loss: 0.984 [71, 240] loss: 0.977 [71, 300] loss: 0.987 [71, 360] loss: 0.970 Epoch: 71 -> Loss: 1.03793370724 Epoch: 71 -> Test Accuracy: 60.14 [72, 60] loss: 0.962 [72, 120] loss: 0.961 [72, 180] loss: 0.965 [72, 240] loss: 0.972 [72, 300] loss: 0.973 [72, 360] loss: 0.975 Epoch: 72 -> Loss: 0.926759541035 Epoch: 72 -> Test Accuracy: 61.0 [73, 60] loss: 0.973 [73, 120] loss: 0.963 [73, 180] loss: 0.976 [73, 240] loss: 0.944 [73, 300] loss: 0.962 [73, 360] loss: 0.968 Epoch: 73 -> Loss: 1.03633105755 Epoch: 73 -> Test Accuracy: 60.63 [74, 60] loss: 0.957 [74, 120] loss: 0.953 [74, 180] loss: 0.968 [74, 240] loss: 0.938 [74, 300] loss: 0.959 [74, 360] loss: 0.964 Epoch: 74 -> Loss: 0.968690276146 Epoch: 74 -> Test Accuracy: 60.56 [75, 60] loss: 0.942 [75, 120] loss: 0.962 [75, 180] loss: 0.942 [75, 240] loss: 0.951 [75, 300] loss: 0.956 [75, 360] loss: 0.946 Epoch: 75 -> Loss: 0.793680846691 Epoch: 75 -> Test Accuracy: 60.38 [76, 60] loss: 0.930 [76, 120] loss: 0.953 [76, 180] loss: 0.942 [76, 240] loss: 0.952 [76, 300] loss: 0.960 [76, 360] loss: 0.956 Epoch: 76 -> Loss: 0.90732383728 Epoch: 76 -> Test Accuracy: 60.94 [77, 60] loss: 0.943 [77, 120] loss: 0.951 [77, 180] loss: 0.960 [77, 240] loss: 0.955 [77, 300] loss: 0.956 [77, 360] loss: 0.943 Epoch: 77 -> Loss: 0.98617619276 Epoch: 77 -> Test Accuracy: 61.15 [78, 60] loss: 0.948 [78, 120] loss: 0.938 [78, 180] loss: 0.961 [78, 240] loss: 0.957 [78, 300] loss: 0.959 [78, 360] loss: 0.946 Epoch: 78 -> Loss: 0.999989330769 Epoch: 78 -> Test Accuracy: 60.75 [79, 60] loss: 0.918 [79, 120] loss: 0.954 [79, 180] loss: 0.962 [79, 240] loss: 0.949 [79, 300] loss: 0.948 [79, 360] loss: 0.980 Epoch: 79 -> Loss: 0.876459002495 Epoch: 79 -> Test Accuracy: 60.38 [80, 60] loss: 0.943 [80, 120] loss: 0.929 [80, 180] loss: 0.947 [80, 240] loss: 0.979 [80, 300] loss: 0.953 [80, 360] loss: 0.958 Epoch: 80 -> Loss: 0.940453529358 Epoch: 80 -> Test Accuracy: 60.24 [81, 60] loss: 0.949 [81, 120] loss: 0.926 [81, 180] loss: 0.955 [81, 240] loss: 0.954 [81, 300] loss: 0.939 [81, 360] loss: 0.953 Epoch: 81 -> Loss: 1.0682592392 Epoch: 81 -> Test Accuracy: 60.5 [82, 60] loss: 0.945 [82, 120] loss: 0.936 [82, 180] loss: 0.953 [82, 240] loss: 0.941 [82, 300] loss: 0.965 [82, 360] loss: 0.955 Epoch: 82 -> Loss: 1.0175538063 Epoch: 82 -> Test Accuracy: 60.48 [83, 60] loss: 0.945 [83, 120] loss: 0.948 [83, 180] loss: 0.934 [83, 240] loss: 0.957 [83, 300] loss: 0.932 [83, 360] loss: 0.948 Epoch: 83 -> Loss: 0.960919976234 Epoch: 83 -> Test Accuracy: 60.91 [84, 60] loss: 0.947 [84, 120] loss: 0.954 [84, 180] loss: 0.942 [84, 240] loss: 0.930 [84, 300] loss: 0.935 [84, 360] loss: 0.949 Epoch: 84 -> Loss: 0.96836745739 Epoch: 84 -> Test Accuracy: 60.92 [85, 60] loss: 0.941 [85, 120] loss: 0.924 [85, 180] loss: 0.941 [85, 240] loss: 0.969 [85, 300] loss: 0.931 [85, 360] loss: 0.955 Epoch: 85 -> Loss: 0.838882446289 Epoch: 85 -> Test Accuracy: 60.38 [86, 60] loss: 0.918 [86, 120] loss: 0.921 [86, 180] loss: 0.934 [86, 240] loss: 0.923 [86, 300] loss: 0.895 [86, 360] loss: 0.918 Epoch: 86 -> Loss: 0.973382174969 Epoch: 86 -> Test Accuracy: 61.58 [87, 60] loss: 0.917 [87, 120] loss: 0.908 [87, 180] loss: 0.925 [87, 240] loss: 0.900 [87, 300] loss: 0.900 [87, 360] loss: 0.899 Epoch: 87 -> Loss: 0.78806501627 Epoch: 87 -> Test Accuracy: 61.5 [88, 60] loss: 0.900 [88, 120] loss: 0.884 [88, 180] loss: 0.924 [88, 240] loss: 0.902 [88, 300] loss: 0.883 [88, 360] loss: 0.911 Epoch: 88 -> Loss: 1.02360868454 Epoch: 88 -> Test Accuracy: 61.6 [89, 60] loss: 0.908 [89, 120] loss: 0.915 [89, 180] loss: 0.895 [89, 240] loss: 0.903 [89, 300] loss: 0.900 [89, 360] loss: 0.905 Epoch: 89 -> Loss: 0.874146580696 Epoch: 89 -> Test Accuracy: 61.86 [90, 60] loss: 0.884 [90, 120] loss: 0.894 [90, 180] loss: 0.901 [90, 240] loss: 0.902 [90, 300] loss: 0.903 [90, 360] loss: 0.893 Epoch: 90 -> Loss: 1.05563819408 Epoch: 90 -> Test Accuracy: 61.69 [91, 60] loss: 0.897 [91, 120] loss: 0.905 [91, 180] loss: 0.898 [91, 240] loss: 0.892 [91, 300] loss: 0.913 [91, 360] loss: 0.901 Epoch: 91 -> Loss: 0.829767525196 Epoch: 91 -> Test Accuracy: 61.99 [92, 60] loss: 0.894 [92, 120] loss: 0.921 [92, 180] loss: 0.899 [92, 240] loss: 0.905 [92, 300] loss: 0.905 [92, 360] loss: 0.893 Epoch: 92 -> Loss: 1.0015989542 Epoch: 92 -> Test Accuracy: 61.87 [93, 60] loss: 0.892 [93, 120] loss: 0.907 [93, 180] loss: 0.910 [93, 240] loss: 0.913 [93, 300] loss: 0.886 [93, 360] loss: 0.907 Epoch: 93 -> Loss: 0.93932056427 Epoch: 93 -> Test Accuracy: 62.14 [94, 60] loss: 0.889 [94, 120] loss: 0.920 [94, 180] loss: 0.905 [94, 240] loss: 0.899 [94, 300] loss: 0.880 [94, 360] loss: 0.887 Epoch: 94 -> Loss: 1.04583060741 Epoch: 94 -> Test Accuracy: 61.96 [95, 60] loss: 0.905 [95, 120] loss: 0.894 [95, 180] loss: 0.888 [95, 240] loss: 0.920 [95, 300] loss: 0.913 [95, 360] loss: 0.884 Epoch: 95 -> Loss: 0.791828036308 Epoch: 95 -> Test Accuracy: 61.88 [96, 60] loss: 0.888 [96, 120] loss: 0.893 [96, 180] loss: 0.903 [96, 240] loss: 0.905 [96, 300] loss: 0.891 [96, 360] loss: 0.889 Epoch: 96 -> Loss: 0.908858656883 Epoch: 96 -> Test Accuracy: 62.08 [97, 60] loss: 0.875 [97, 120] loss: 0.908 [97, 180] loss: 0.913 [97, 240] loss: 0.902 [97, 300] loss: 0.892 [97, 360] loss: 0.905 Epoch: 97 -> Loss: 0.757797002792 Epoch: 97 -> Test Accuracy: 61.86 [98, 60] loss: 0.886 [98, 120] loss: 0.902 [98, 180] loss: 0.905 [98, 240] loss: 0.897 [98, 300] loss: 0.903 [98, 360] loss: 0.882 Epoch: 98 -> Loss: 0.870647132397 Epoch: 98 -> Test Accuracy: 61.84 [99, 60] loss: 0.908 [99, 120] loss: 0.892 [99, 180] loss: 0.893 [99, 240] loss: 0.900 [99, 300] loss: 0.889 [99, 360] loss: 0.883 Epoch: 99 -> Loss: 0.933106899261 Epoch: 99 -> Test Accuracy: 62.03 [100, 60] loss: 0.890 [100, 120] loss: 0.895 [100, 180] loss: 0.881 [100, 240] loss: 0.893 [100, 300] loss: 0.891 [100, 360] loss: 0.896 Epoch: 100 -> Loss: 0.887984871864 Epoch: 100 -> Test Accuracy: 61.91 Finished Training
In [16]:
# save variables
fm.save_variable([rot_block3_loss_log, rot_block3_test_accuracy_log,
block3_loss_log, block3_test_accuracy_log,
conv_block3_loss_log, conv_block3_test_accuracy_log], "3_block_net")
In [17]:
# rename files
fm.add_block_to_name(3, [100, 200])
4 Block RotNet¶
In [6]:
# initialize network
net_block4 = RN.RotNet(num_classes=4, num_conv_block=4, add_avg_pool=False)
In [7]:
# train network
rot_block4_loss_log, _, rot_block4_test_accuracy_log, _, _ = tr.adaptive_learning([0.1, 0.02, 0.004, 0.0008],
[60, 120, 160, 200], 0.9, 5e-4, net_block4, criterion, trainloader, None, testloader, rot=['90', '180', '270'])
functionalities/rotation.py:16: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. flip_idx = torch.range(trans_im.size(2) - 1, 0, -1).long() functionalities/rotation.py:31: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. vert_idx = torch.range(image.size(2) - 1, 0, -1).long() functionalities/rotation.py:35: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. hor_idx = torch.range(vert_im.size(1) - 1, 0, -1).long() functionalities/rotation.py:50: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. vert_idx = torch.range(image.size(2) - 1, 0, -1).long()
[1, 60] loss: 1.280 [1, 120] loss: 1.013 [1, 180] loss: 0.969 [1, 240] loss: 0.922 [1, 300] loss: 0.892 [1, 360] loss: 0.833 Epoch: 1 -> Loss: 0.88242161274 Epoch: 1 -> Test Accuracy: 65.2375 [2, 60] loss: 0.788 [2, 120] loss: 0.754 [2, 180] loss: 0.742 [2, 240] loss: 0.717 [2, 300] loss: 0.698 [2, 360] loss: 0.669 Epoch: 2 -> Loss: 0.673194825649 Epoch: 2 -> Test Accuracy: 73.2225 [3, 60] loss: 0.646 [3, 120] loss: 0.647 [3, 180] loss: 0.610 [3, 240] loss: 0.634 [3, 300] loss: 0.590 [3, 360] loss: 0.590 Epoch: 3 -> Loss: 0.540262639523 Epoch: 3 -> Test Accuracy: 75.73 [4, 60] loss: 0.553 [4, 120] loss: 0.572 [4, 180] loss: 0.568 [4, 240] loss: 0.564 [4, 300] loss: 0.545 [4, 360] loss: 0.531 Epoch: 4 -> Loss: 0.632760047913 Epoch: 4 -> Test Accuracy: 78.83 [5, 60] loss: 0.514 [5, 120] loss: 0.521 [5, 180] loss: 0.504 [5, 240] loss: 0.530 [5, 300] loss: 0.500 [5, 360] loss: 0.511 Epoch: 5 -> Loss: 0.40007519722 Epoch: 5 -> Test Accuracy: 80.075 [6, 60] loss: 0.487 [6, 120] loss: 0.484 [6, 180] loss: 0.494 [6, 240] loss: 0.486 [6, 300] loss: 0.482 [6, 360] loss: 0.475 Epoch: 6 -> Loss: 0.572343051434 Epoch: 6 -> Test Accuracy: 80.91 [7, 60] loss: 0.455 [7, 120] loss: 0.479 [7, 180] loss: 0.460 [7, 240] loss: 0.466 [7, 300] loss: 0.464 [7, 360] loss: 0.454 Epoch: 7 -> Loss: 0.433941841125 Epoch: 7 -> Test Accuracy: 82.3975 [8, 60] loss: 0.441 [8, 120] loss: 0.462 [8, 180] loss: 0.432 [8, 240] loss: 0.440 [8, 300] loss: 0.455 [8, 360] loss: 0.447 Epoch: 8 -> Loss: 0.374135434628 Epoch: 8 -> Test Accuracy: 82.39 [9, 60] loss: 0.426 [9, 120] loss: 0.440 [9, 180] loss: 0.437 [9, 240] loss: 0.436 [9, 300] loss: 0.422 [9, 360] loss: 0.416 Epoch: 9 -> Loss: 0.478541225195 Epoch: 9 -> Test Accuracy: 83.175 [10, 60] loss: 0.410 [10, 120] loss: 0.423 [10, 180] loss: 0.409 [10, 240] loss: 0.428 [10, 300] loss: 0.415 [10, 360] loss: 0.413 Epoch: 10 -> Loss: 0.467094898224 Epoch: 10 -> Test Accuracy: 82.305 [11, 60] loss: 0.405 [11, 120] loss: 0.406 [11, 180] loss: 0.421 [11, 240] loss: 0.403 [11, 300] loss: 0.405 [11, 360] loss: 0.401 Epoch: 11 -> Loss: 0.471219927073 Epoch: 11 -> Test Accuracy: 83.6775 [12, 60] loss: 0.387 [12, 120] loss: 0.400 [12, 180] loss: 0.406 [12, 240] loss: 0.388 [12, 300] loss: 0.382 [12, 360] loss: 0.408 Epoch: 12 -> Loss: 0.430007696152 Epoch: 12 -> Test Accuracy: 83.945 [13, 60] loss: 0.384 [13, 120] loss: 0.387 [13, 180] loss: 0.377 [13, 240] loss: 0.385 [13, 300] loss: 0.380 [13, 360] loss: 0.399 Epoch: 13 -> Loss: 0.338429063559 Epoch: 13 -> Test Accuracy: 84.635 [14, 60] loss: 0.369 [14, 120] loss: 0.384 [14, 180] loss: 0.371 [14, 240] loss: 0.386 [14, 300] loss: 0.378 [14, 360] loss: 0.392 Epoch: 14 -> Loss: 0.285325229168 Epoch: 14 -> Test Accuracy: 83.9125 [15, 60] loss: 0.370 [15, 120] loss: 0.370 [15, 180] loss: 0.380 [15, 240] loss: 0.363 [15, 300] loss: 0.378 [15, 360] loss: 0.376 Epoch: 15 -> Loss: 0.269873142242 Epoch: 15 -> Test Accuracy: 85.4075 [16, 60] loss: 0.351 [16, 120] loss: 0.369 [16, 180] loss: 0.366 [16, 240] loss: 0.367 [16, 300] loss: 0.377 [16, 360] loss: 0.374 Epoch: 16 -> Loss: 0.428826898336 Epoch: 16 -> Test Accuracy: 84.875 [17, 60] loss: 0.357 [17, 120] loss: 0.342 [17, 180] loss: 0.373 [17, 240] loss: 0.355 [17, 300] loss: 0.364 [17, 360] loss: 0.364 Epoch: 17 -> Loss: 0.411879599094 Epoch: 17 -> Test Accuracy: 85.5875 [18, 60] loss: 0.345 [18, 120] loss: 0.349 [18, 180] loss: 0.358 [18, 240] loss: 0.357 [18, 300] loss: 0.364 [18, 360] loss: 0.359 Epoch: 18 -> Loss: 0.339927792549 Epoch: 18 -> Test Accuracy: 85.1625 [19, 60] loss: 0.342 [19, 120] loss: 0.361 [19, 180] loss: 0.350 [19, 240] loss: 0.351 [19, 300] loss: 0.352 [19, 360] loss: 0.370 Epoch: 19 -> Loss: 0.245223641396 Epoch: 19 -> Test Accuracy: 85.005 [20, 60] loss: 0.341 [20, 120] loss: 0.336 [20, 180] loss: 0.338 [20, 240] loss: 0.366 [20, 300] loss: 0.356 [20, 360] loss: 0.352 Epoch: 20 -> Loss: 0.386262714863 Epoch: 20 -> Test Accuracy: 85.4325 [21, 60] loss: 0.341 [21, 120] loss: 0.335 [21, 180] loss: 0.358 [21, 240] loss: 0.350 [21, 300] loss: 0.347 [21, 360] loss: 0.347 Epoch: 21 -> Loss: 0.457043260336 Epoch: 21 -> Test Accuracy: 86.125 [22, 60] loss: 0.332 [22, 120] loss: 0.342 [22, 180] loss: 0.338 [22, 240] loss: 0.333 [22, 300] loss: 0.356 [22, 360] loss: 0.347 Epoch: 22 -> Loss: 0.478622019291 Epoch: 22 -> Test Accuracy: 86.1325 [23, 60] loss: 0.325 [23, 120] loss: 0.333 [23, 180] loss: 0.340 [23, 240] loss: 0.335 [23, 300] loss: 0.336 [23, 360] loss: 0.351 Epoch: 23 -> Loss: 0.37189117074 Epoch: 23 -> Test Accuracy: 86.0975 [24, 60] loss: 0.332 [24, 120] loss: 0.334 [24, 180] loss: 0.350 [24, 240] loss: 0.322 [24, 300] loss: 0.331 [24, 360] loss: 0.342 Epoch: 24 -> Loss: 0.315701901913 Epoch: 24 -> Test Accuracy: 84.5375 [25, 60] loss: 0.325 [25, 120] loss: 0.332 [25, 180] loss: 0.326 [25, 240] loss: 0.338 [25, 300] loss: 0.342 [25, 360] loss: 0.345 Epoch: 25 -> Loss: 0.295279443264 Epoch: 25 -> Test Accuracy: 86.34 [26, 60] loss: 0.334 [26, 120] loss: 0.326 [26, 180] loss: 0.325 [26, 240] loss: 0.348 [26, 300] loss: 0.331 [26, 360] loss: 0.332 Epoch: 26 -> Loss: 0.388910293579 Epoch: 26 -> Test Accuracy: 85.57 [27, 60] loss: 0.315 [27, 120] loss: 0.327 [27, 180] loss: 0.334 [27, 240] loss: 0.323 [27, 300] loss: 0.337 [27, 360] loss: 0.329 Epoch: 27 -> Loss: 0.313533097506 Epoch: 27 -> Test Accuracy: 85.2975 [28, 60] loss: 0.310 [28, 120] loss: 0.329 [28, 180] loss: 0.321 [28, 240] loss: 0.330 [28, 300] loss: 0.340 [28, 360] loss: 0.330 Epoch: 28 -> Loss: 0.30606174469 Epoch: 28 -> Test Accuracy: 86.3625 [29, 60] loss: 0.311 [29, 120] loss: 0.319 [29, 180] loss: 0.329 [29, 240] loss: 0.326 [29, 300] loss: 0.331 [29, 360] loss: 0.335 Epoch: 29 -> Loss: 0.292320221663 Epoch: 29 -> Test Accuracy: 86.36 [30, 60] loss: 0.304 [30, 120] loss: 0.323 [30, 180] loss: 0.326 [30, 240] loss: 0.323 [30, 300] loss: 0.311 [30, 360] loss: 0.326 Epoch: 30 -> Loss: 0.377558887005 Epoch: 30 -> Test Accuracy: 85.0525 [31, 60] loss: 0.323 [31, 120] loss: 0.319 [31, 180] loss: 0.314 [31, 240] loss: 0.307 [31, 300] loss: 0.331 [31, 360] loss: 0.324 Epoch: 31 -> Loss: 0.309758841991 Epoch: 31 -> Test Accuracy: 86.3675 [32, 60] loss: 0.301 [32, 120] loss: 0.313 [32, 180] loss: 0.326 [32, 240] loss: 0.313 [32, 300] loss: 0.326 [32, 360] loss: 0.328 Epoch: 32 -> Loss: 0.241686820984 Epoch: 32 -> Test Accuracy: 85.84 [33, 60] loss: 0.316 [33, 120] loss: 0.313 [33, 180] loss: 0.312 [33, 240] loss: 0.325 [33, 300] loss: 0.319 [33, 360] loss: 0.323 Epoch: 33 -> Loss: 0.333684593439 Epoch: 33 -> Test Accuracy: 86.545 [34, 60] loss: 0.312 [34, 120] loss: 0.312 [34, 180] loss: 0.322 [34, 240] loss: 0.321 [34, 300] loss: 0.322 [34, 360] loss: 0.311 Epoch: 34 -> Loss: 0.388200581074 Epoch: 34 -> Test Accuracy: 85.8125 [35, 60] loss: 0.312 [35, 120] loss: 0.307 [35, 180] loss: 0.321 [35, 240] loss: 0.313 [35, 300] loss: 0.323 [35, 360] loss: 0.323 Epoch: 35 -> Loss: 0.271653354168 Epoch: 35 -> Test Accuracy: 86.68 [36, 60] loss: 0.299 [36, 120] loss: 0.304 [36, 180] loss: 0.314 [36, 240] loss: 0.329 [36, 300] loss: 0.316 [36, 360] loss: 0.326 Epoch: 36 -> Loss: 0.332448899746 Epoch: 36 -> Test Accuracy: 85.805 [37, 60] loss: 0.300 [37, 120] loss: 0.307 [37, 180] loss: 0.316 [37, 240] loss: 0.315 [37, 300] loss: 0.318 [37, 360] loss: 0.312 Epoch: 37 -> Loss: 0.253511637449 Epoch: 37 -> Test Accuracy: 86.8125 [38, 60] loss: 0.302 [38, 120] loss: 0.314 [38, 180] loss: 0.312 [38, 240] loss: 0.315 [38, 300] loss: 0.313 [38, 360] loss: 0.309 Epoch: 38 -> Loss: 0.288749873638 Epoch: 38 -> Test Accuracy: 86.67 [39, 60] loss: 0.292 [39, 120] loss: 0.321 [39, 180] loss: 0.322 [39, 240] loss: 0.305 [39, 300] loss: 0.308 [39, 360] loss: 0.330 Epoch: 39 -> Loss: 0.351736664772 Epoch: 39 -> Test Accuracy: 86.85 [40, 60] loss: 0.297 [40, 120] loss: 0.315 [40, 180] loss: 0.308 [40, 240] loss: 0.321 [40, 300] loss: 0.308 [40, 360] loss: 0.323 Epoch: 40 -> Loss: 0.299942553043 Epoch: 40 -> Test Accuracy: 86.61 [41, 60] loss: 0.301 [41, 120] loss: 0.305 [41, 180] loss: 0.295 [41, 240] loss: 0.319 [41, 300] loss: 0.305 [41, 360] loss: 0.316 Epoch: 41 -> Loss: 0.281109631062 Epoch: 41 -> Test Accuracy: 87.31 [42, 60] loss: 0.290 [42, 120] loss: 0.320 [42, 180] loss: 0.308 [42, 240] loss: 0.302 [42, 300] loss: 0.311 [42, 360] loss: 0.311 Epoch: 42 -> Loss: 0.37639850378 Epoch: 42 -> Test Accuracy: 86.91 [43, 60] loss: 0.300 [43, 120] loss: 0.310 [43, 180] loss: 0.313 [43, 240] loss: 0.310 [43, 300] loss: 0.307 [43, 360] loss: 0.306 Epoch: 43 -> Loss: 0.322246402502 Epoch: 43 -> Test Accuracy: 85.7625 [44, 60] loss: 0.300 [44, 120] loss: 0.294 [44, 180] loss: 0.306 [44, 240] loss: 0.317 [44, 300] loss: 0.324 [44, 360] loss: 0.302 Epoch: 44 -> Loss: 0.329386383295 Epoch: 44 -> Test Accuracy: 86.535 [45, 60] loss: 0.285 [45, 120] loss: 0.302 [45, 180] loss: 0.310 [45, 240] loss: 0.303 [45, 300] loss: 0.316 [45, 360] loss: 0.320 Epoch: 45 -> Loss: 0.343748152256 Epoch: 45 -> Test Accuracy: 86.4875 [46, 60] loss: 0.297 [46, 120] loss: 0.300 [46, 180] loss: 0.307 [46, 240] loss: 0.305 [46, 300] loss: 0.321 [46, 360] loss: 0.317 Epoch: 46 -> Loss: 0.241001099348 Epoch: 46 -> Test Accuracy: 87.6325 [47, 60] loss: 0.287 [47, 120] loss: 0.305 [47, 180] loss: 0.306 [47, 240] loss: 0.311 [47, 300] loss: 0.309 [47, 360] loss: 0.313 Epoch: 47 -> Loss: 0.264798223972 Epoch: 47 -> Test Accuracy: 86.7775 [48, 60] loss: 0.292 [48, 120] loss: 0.300 [48, 180] loss: 0.297 [48, 240] loss: 0.308 [48, 300] loss: 0.299 [48, 360] loss: 0.315 Epoch: 48 -> Loss: 0.321404635906 Epoch: 48 -> Test Accuracy: 86.86 [49, 60] loss: 0.291 [49, 120] loss: 0.304 [49, 180] loss: 0.294 [49, 240] loss: 0.318 [49, 300] loss: 0.318 [49, 360] loss: 0.298 Epoch: 49 -> Loss: 0.368998676538 Epoch: 49 -> Test Accuracy: 87.6475 [50, 60] loss: 0.304 [50, 120] loss: 0.302 [50, 180] loss: 0.299 [50, 240] loss: 0.311 [50, 300] loss: 0.300 [50, 360] loss: 0.313 Epoch: 50 -> Loss: 0.368564456701 Epoch: 50 -> Test Accuracy: 86.975 [51, 60] loss: 0.289 [51, 120] loss: 0.308 [51, 180] loss: 0.282 [51, 240] loss: 0.312 [51, 300] loss: 0.299 [51, 360] loss: 0.306 Epoch: 51 -> Loss: 0.376721441746 Epoch: 51 -> Test Accuracy: 85.77 [52, 60] loss: 0.283 [52, 120] loss: 0.298 [52, 180] loss: 0.300 [52, 240] loss: 0.305 [52, 300] loss: 0.298 [52, 360] loss: 0.300 Epoch: 52 -> Loss: 0.283972710371 Epoch: 52 -> Test Accuracy: 86.7825 [53, 60] loss: 0.286 [53, 120] loss: 0.298 [53, 180] loss: 0.291 [53, 240] loss: 0.299 [53, 300] loss: 0.301 [53, 360] loss: 0.304 Epoch: 53 -> Loss: 0.31412217021 Epoch: 53 -> Test Accuracy: 86.3825 [54, 60] loss: 0.298 [54, 120] loss: 0.286 [54, 180] loss: 0.312 [54, 240] loss: 0.295 [54, 300] loss: 0.292 [54, 360] loss: 0.301 Epoch: 54 -> Loss: 0.356736570597 Epoch: 54 -> Test Accuracy: 86.355 [55, 60] loss: 0.290 [55, 120] loss: 0.295 [55, 180] loss: 0.308 [55, 240] loss: 0.304 [55, 300] loss: 0.303 [55, 360] loss: 0.301 Epoch: 55 -> Loss: 0.331777065992 Epoch: 55 -> Test Accuracy: 86.665 [56, 60] loss: 0.289 [56, 120] loss: 0.297 [56, 180] loss: 0.299 [56, 240] loss: 0.307 [56, 300] loss: 0.287 [56, 360] loss: 0.321 Epoch: 56 -> Loss: 0.261130779982 Epoch: 56 -> Test Accuracy: 86.6275 [57, 60] loss: 0.280 [57, 120] loss: 0.289 [57, 180] loss: 0.309 [57, 240] loss: 0.309 [57, 300] loss: 0.301 [57, 360] loss: 0.302 Epoch: 57 -> Loss: 0.289793878794 Epoch: 57 -> Test Accuracy: 86.805 [58, 60] loss: 0.292 [58, 120] loss: 0.300 [58, 180] loss: 0.300 [58, 240] loss: 0.301 [58, 300] loss: 0.294 [58, 360] loss: 0.308 Epoch: 58 -> Loss: 0.391004383564 Epoch: 58 -> Test Accuracy: 86.545 [59, 60] loss: 0.299 [59, 120] loss: 0.292 [59, 180] loss: 0.300 [59, 240] loss: 0.288 [59, 300] loss: 0.304 [59, 360] loss: 0.296 Epoch: 59 -> Loss: 0.291617512703 Epoch: 59 -> Test Accuracy: 86.95 [60, 60] loss: 0.282 [60, 120] loss: 0.299 [60, 180] loss: 0.304 [60, 240] loss: 0.310 [60, 300] loss: 0.299 [60, 360] loss: 0.298 Epoch: 60 -> Loss: 0.328978180885 Epoch: 60 -> Test Accuracy: 87.535 [61, 60] loss: 0.225 [61, 120] loss: 0.194 [61, 180] loss: 0.181 [61, 240] loss: 0.183 [61, 300] loss: 0.183 [61, 360] loss: 0.183 Epoch: 61 -> Loss: 0.126734361053 Epoch: 61 -> Test Accuracy: 91.32 [62, 60] loss: 0.158 [62, 120] loss: 0.152 [62, 180] loss: 0.155 [62, 240] loss: 0.166 [62, 300] loss: 0.176 [62, 360] loss: 0.172 Epoch: 62 -> Loss: 0.174959421158 Epoch: 62 -> Test Accuracy: 91.5325 [63, 60] loss: 0.149 [63, 120] loss: 0.152 [63, 180] loss: 0.158 [63, 240] loss: 0.160 [63, 300] loss: 0.161 [63, 360] loss: 0.156 Epoch: 63 -> Loss: 0.209470033646 Epoch: 63 -> Test Accuracy: 91.5125 [64, 60] loss: 0.138 [64, 120] loss: 0.148 [64, 180] loss: 0.152 [64, 240] loss: 0.152 [64, 300] loss: 0.164 [64, 360] loss: 0.157 Epoch: 64 -> Loss: 0.136709198356 Epoch: 64 -> Test Accuracy: 91.1275 [65, 60] loss: 0.133 [65, 120] loss: 0.142 [65, 180] loss: 0.155 [65, 240] loss: 0.146 [65, 300] loss: 0.152 [65, 360] loss: 0.152 Epoch: 65 -> Loss: 0.118245780468 Epoch: 65 -> Test Accuracy: 91.0175 [66, 60] loss: 0.133 [66, 120] loss: 0.132 [66, 180] loss: 0.145 [66, 240] loss: 0.138 [66, 300] loss: 0.152 [66, 360] loss: 0.150 Epoch: 66 -> Loss: 0.106802843511 Epoch: 66 -> Test Accuracy: 90.7525 [67, 60] loss: 0.141 [67, 120] loss: 0.142 [67, 180] loss: 0.141 [67, 240] loss: 0.146 [67, 300] loss: 0.152 [67, 360] loss: 0.142 Epoch: 67 -> Loss: 0.183747336268 Epoch: 67 -> Test Accuracy: 90.9525 [68, 60] loss: 0.137 [68, 120] loss: 0.135 [68, 180] loss: 0.155 [68, 240] loss: 0.142 [68, 300] loss: 0.155 [68, 360] loss: 0.144 Epoch: 68 -> Loss: 0.127017647028 Epoch: 68 -> Test Accuracy: 90.4275 [69, 60] loss: 0.133 [69, 120] loss: 0.134 [69, 180] loss: 0.136 [69, 240] loss: 0.153 [69, 300] loss: 0.151 [69, 360] loss: 0.152 Epoch: 69 -> Loss: 0.146195858717 Epoch: 69 -> Test Accuracy: 90.5875 [70, 60] loss: 0.141 [70, 120] loss: 0.136 [70, 180] loss: 0.140 [70, 240] loss: 0.147 [70, 300] loss: 0.152 [70, 360] loss: 0.145 Epoch: 70 -> Loss: 0.144819706678 Epoch: 70 -> Test Accuracy: 90.445 [71, 60] loss: 0.133 [71, 120] loss: 0.143 [71, 180] loss: 0.135 [71, 240] loss: 0.152 [71, 300] loss: 0.149 [71, 360] loss: 0.158 Epoch: 71 -> Loss: 0.122240677476 Epoch: 71 -> Test Accuracy: 90.6475 [72, 60] loss: 0.133 [72, 120] loss: 0.139 [72, 180] loss: 0.146 [72, 240] loss: 0.144 [72, 300] loss: 0.149 [72, 360] loss: 0.152 Epoch: 72 -> Loss: 0.121262550354 Epoch: 72 -> Test Accuracy: 90.77 [73, 60] loss: 0.131 [73, 120] loss: 0.142 [73, 180] loss: 0.141 [73, 240] loss: 0.152 [73, 300] loss: 0.151 [73, 360] loss: 0.162 Epoch: 73 -> Loss: 0.174145117402 Epoch: 73 -> Test Accuracy: 90.6425 [74, 60] loss: 0.134 [74, 120] loss: 0.138 [74, 180] loss: 0.157 [74, 240] loss: 0.148 [74, 300] loss: 0.141 [74, 360] loss: 0.154 Epoch: 74 -> Loss: 0.0880781561136 Epoch: 74 -> Test Accuracy: 91.04 [75, 60] loss: 0.138 [75, 120] loss: 0.141 [75, 180] loss: 0.149 [75, 240] loss: 0.146 [75, 300] loss: 0.147 [75, 360] loss: 0.142 Epoch: 75 -> Loss: 0.126393944025 Epoch: 75 -> Test Accuracy: 90.82 [76, 60] loss: 0.134 [76, 120] loss: 0.136 [76, 180] loss: 0.136 [76, 240] loss: 0.157 [76, 300] loss: 0.148 [76, 360] loss: 0.150 Epoch: 76 -> Loss: 0.113145872951 Epoch: 76 -> Test Accuracy: 90.26 [77, 60] loss: 0.137 [77, 120] loss: 0.144 [77, 180] loss: 0.143 [77, 240] loss: 0.144 [77, 300] loss: 0.157 [77, 360] loss: 0.154 Epoch: 77 -> Loss: 0.189672544599 Epoch: 77 -> Test Accuracy: 90.6525 [78, 60] loss: 0.132 [78, 120] loss: 0.145 [78, 180] loss: 0.145 [78, 240] loss: 0.145 [78, 300] loss: 0.154 [78, 360] loss: 0.157 Epoch: 78 -> Loss: 0.134117081761 Epoch: 78 -> Test Accuracy: 90.925 [79, 60] loss: 0.136 [79, 120] loss: 0.133 [79, 180] loss: 0.146 [79, 240] loss: 0.144 [79, 300] loss: 0.153 [79, 360] loss: 0.155 Epoch: 79 -> Loss: 0.178136751056 Epoch: 79 -> Test Accuracy: 90.68 [80, 60] loss: 0.128 [80, 120] loss: 0.138 [80, 180] loss: 0.143 [80, 240] loss: 0.147 [80, 300] loss: 0.151 [80, 360] loss: 0.141 Epoch: 80 -> Loss: 0.193272918463 Epoch: 80 -> Test Accuracy: 90.4525 [81, 60] loss: 0.135 [81, 120] loss: 0.132 [81, 180] loss: 0.142 [81, 240] loss: 0.151 [81, 300] loss: 0.156 [81, 360] loss: 0.160 Epoch: 81 -> Loss: 0.140550598502 Epoch: 81 -> Test Accuracy: 90.345 [82, 60] loss: 0.136 [82, 120] loss: 0.140 [82, 180] loss: 0.139 [82, 240] loss: 0.152 [82, 300] loss: 0.152 [82, 360] loss: 0.154 Epoch: 82 -> Loss: 0.1007431373 Epoch: 82 -> Test Accuracy: 90.8 [83, 60] loss: 0.130 [83, 120] loss: 0.140 [83, 180] loss: 0.145 [83, 240] loss: 0.151 [83, 300] loss: 0.153 [83, 360] loss: 0.147 Epoch: 83 -> Loss: 0.162469476461 Epoch: 83 -> Test Accuracy: 90.6975 [84, 60] loss: 0.136 [84, 120] loss: 0.144 [84, 180] loss: 0.149 [84, 240] loss: 0.134 [84, 300] loss: 0.143 [84, 360] loss: 0.154 Epoch: 84 -> Loss: 0.161104053259 Epoch: 84 -> Test Accuracy: 90.5925 [85, 60] loss: 0.136 [85, 120] loss: 0.138 [85, 180] loss: 0.153 [85, 240] loss: 0.150 [85, 300] loss: 0.145 [85, 360] loss: 0.144 Epoch: 85 -> Loss: 0.124722383916 Epoch: 85 -> Test Accuracy: 90.1225 [86, 60] loss: 0.130 [86, 120] loss: 0.137 [86, 180] loss: 0.140 [86, 240] loss: 0.153 [86, 300] loss: 0.148 [86, 360] loss: 0.153 Epoch: 86 -> Loss: 0.141249030828 Epoch: 86 -> Test Accuracy: 90.0975 [87, 60] loss: 0.129 [87, 120] loss: 0.140 [87, 180] loss: 0.144 [87, 240] loss: 0.142 [87, 300] loss: 0.146 [87, 360] loss: 0.150 Epoch: 87 -> Loss: 0.164696201682 Epoch: 87 -> Test Accuracy: 90.755 [88, 60] loss: 0.133 [88, 120] loss: 0.135 [88, 180] loss: 0.150 [88, 240] loss: 0.139 [88, 300] loss: 0.149 [88, 360] loss: 0.156 Epoch: 88 -> Loss: 0.12807802856 Epoch: 88 -> Test Accuracy: 90.385 [89, 60] loss: 0.138 [89, 120] loss: 0.138 [89, 180] loss: 0.135 [89, 240] loss: 0.146 [89, 300] loss: 0.142 [89, 360] loss: 0.143 Epoch: 89 -> Loss: 0.131286710501 Epoch: 89 -> Test Accuracy: 90.56 [90, 60] loss: 0.129 [90, 120] loss: 0.134 [90, 180] loss: 0.138 [90, 240] loss: 0.146 [90, 300] loss: 0.151 [90, 360] loss: 0.154 Epoch: 90 -> Loss: 0.173849195242 Epoch: 90 -> Test Accuracy: 90.23 [91, 60] loss: 0.129 [91, 120] loss: 0.132 [91, 180] loss: 0.149 [91, 240] loss: 0.139 [91, 300] loss: 0.145 [91, 360] loss: 0.152 Epoch: 91 -> Loss: 0.149797052145 Epoch: 91 -> Test Accuracy: 89.45 [92, 60] loss: 0.128 [92, 120] loss: 0.137 [92, 180] loss: 0.137 [92, 240] loss: 0.154 [92, 300] loss: 0.147 [92, 360] loss: 0.141 Epoch: 92 -> Loss: 0.115692041814 Epoch: 92 -> Test Accuracy: 90.2425 [93, 60] loss: 0.124 [93, 120] loss: 0.134 [93, 180] loss: 0.131 [93, 240] loss: 0.140 [93, 300] loss: 0.145 [93, 360] loss: 0.153 Epoch: 93 -> Loss: 0.165478780866 Epoch: 93 -> Test Accuracy: 90.62 [94, 60] loss: 0.141 [94, 120] loss: 0.127 [94, 180] loss: 0.134 [94, 240] loss: 0.133 [94, 300] loss: 0.151 [94, 360] loss: 0.148 Epoch: 94 -> Loss: 0.180439129472 Epoch: 94 -> Test Accuracy: 90.4225 [95, 60] loss: 0.128 [95, 120] loss: 0.132 [95, 180] loss: 0.141 [95, 240] loss: 0.137 [95, 300] loss: 0.139 [95, 360] loss: 0.147 Epoch: 95 -> Loss: 0.18649956584 Epoch: 95 -> Test Accuracy: 90.5 [96, 60] loss: 0.132 [96, 120] loss: 0.132 [96, 180] loss: 0.134 [96, 240] loss: 0.138 [96, 300] loss: 0.148 [96, 360] loss: 0.137 Epoch: 96 -> Loss: 0.127098694444 Epoch: 96 -> Test Accuracy: 90.81 [97, 60] loss: 0.127 [97, 120] loss: 0.132 [97, 180] loss: 0.133 [97, 240] loss: 0.150 [97, 300] loss: 0.141 [97, 360] loss: 0.147 Epoch: 97 -> Loss: 0.130502581596 Epoch: 97 -> Test Accuracy: 90.665 [98, 60] loss: 0.131 [98, 120] loss: 0.132 [98, 180] loss: 0.135 [98, 240] loss: 0.144 [98, 300] loss: 0.138 [98, 360] loss: 0.140 Epoch: 98 -> Loss: 0.194058328867 Epoch: 98 -> Test Accuracy: 90.675 [99, 60] loss: 0.126 [99, 120] loss: 0.124 [99, 180] loss: 0.145 [99, 240] loss: 0.133 [99, 300] loss: 0.141 [99, 360] loss: 0.141 Epoch: 99 -> Loss: 0.0897030010819 Epoch: 99 -> Test Accuracy: 90.155 [100, 60] loss: 0.133 [100, 120] loss: 0.128 [100, 180] loss: 0.133 [100, 240] loss: 0.139 [100, 300] loss: 0.140 [100, 360] loss: 0.142 Epoch: 100 -> Loss: 0.149182528257 Epoch: 100 -> Test Accuracy: 90.305 [101, 60] loss: 0.124 [101, 120] loss: 0.135 [101, 180] loss: 0.132 [101, 240] loss: 0.138 [101, 300] loss: 0.143 [101, 360] loss: 0.134 Epoch: 101 -> Loss: 0.196269705892 Epoch: 101 -> Test Accuracy: 90.4925 [102, 60] loss: 0.123 [102, 120] loss: 0.130 [102, 180] loss: 0.134 [102, 240] loss: 0.142 [102, 300] loss: 0.140 [102, 360] loss: 0.139 Epoch: 102 -> Loss: 0.164475351572 Epoch: 102 -> Test Accuracy: 90.125 [103, 60] loss: 0.131 [103, 120] loss: 0.132 [103, 180] loss: 0.137 [103, 240] loss: 0.129 [103, 300] loss: 0.143 [103, 360] loss: 0.138 Epoch: 103 -> Loss: 0.356296956539 Epoch: 103 -> Test Accuracy: 90.2525 [104, 60] loss: 0.120 [104, 120] loss: 0.133 [104, 180] loss: 0.131 [104, 240] loss: 0.134 [104, 300] loss: 0.143 [104, 360] loss: 0.137 Epoch: 104 -> Loss: 0.180964976549 Epoch: 104 -> Test Accuracy: 90.5475 [105, 60] loss: 0.125 [105, 120] loss: 0.131 [105, 180] loss: 0.138 [105, 240] loss: 0.133 [105, 300] loss: 0.126 [105, 360] loss: 0.139 Epoch: 105 -> Loss: 0.1654535532 Epoch: 105 -> Test Accuracy: 90.33 [106, 60] loss: 0.127 [106, 120] loss: 0.129 [106, 180] loss: 0.133 [106, 240] loss: 0.138 [106, 300] loss: 0.137 [106, 360] loss: 0.150 Epoch: 106 -> Loss: 0.110846780241 Epoch: 106 -> Test Accuracy: 90.7225 [107, 60] loss: 0.131 [107, 120] loss: 0.135 [107, 180] loss: 0.134 [107, 240] loss: 0.136 [107, 300] loss: 0.133 [107, 360] loss: 0.146 Epoch: 107 -> Loss: 0.0828087627888 Epoch: 107 -> Test Accuracy: 90.3075 [108, 60] loss: 0.124 [108, 120] loss: 0.122 [108, 180] loss: 0.132 [108, 240] loss: 0.130 [108, 300] loss: 0.139 [108, 360] loss: 0.139 Epoch: 108 -> Loss: 0.22273555398 Epoch: 108 -> Test Accuracy: 90.74 [109, 60] loss: 0.128 [109, 120] loss: 0.128 [109, 180] loss: 0.130 [109, 240] loss: 0.144 [109, 300] loss: 0.140 [109, 360] loss: 0.139 Epoch: 109 -> Loss: 0.138728111982 Epoch: 109 -> Test Accuracy: 90.34 [110, 60] loss: 0.115 [110, 120] loss: 0.134 [110, 180] loss: 0.127 [110, 240] loss: 0.129 [110, 300] loss: 0.138 [110, 360] loss: 0.138 Epoch: 110 -> Loss: 0.154812350869 Epoch: 110 -> Test Accuracy: 90.495 [111, 60] loss: 0.124 [111, 120] loss: 0.127 [111, 180] loss: 0.130 [111, 240] loss: 0.136 [111, 300] loss: 0.133 [111, 360] loss: 0.135 Epoch: 111 -> Loss: 0.180580988526 Epoch: 111 -> Test Accuracy: 89.7225 [112, 60] loss: 0.136 [112, 120] loss: 0.125 [112, 180] loss: 0.132 [112, 240] loss: 0.130 [112, 300] loss: 0.137 [112, 360] loss: 0.140 Epoch: 112 -> Loss: 0.158955469728 Epoch: 112 -> Test Accuracy: 90.2075 [113, 60] loss: 0.127 [113, 120] loss: 0.122 [113, 180] loss: 0.133 [113, 240] loss: 0.138 [113, 300] loss: 0.134 [113, 360] loss: 0.139 Epoch: 113 -> Loss: 0.124354101717 Epoch: 113 -> Test Accuracy: 90.2075 [114, 60] loss: 0.124 [114, 120] loss: 0.121 [114, 180] loss: 0.133 [114, 240] loss: 0.126 [114, 300] loss: 0.137 [114, 360] loss: 0.140 Epoch: 114 -> Loss: 0.117385149002 Epoch: 114 -> Test Accuracy: 90.535 [115, 60] loss: 0.125 [115, 120] loss: 0.132 [115, 180] loss: 0.135 [115, 240] loss: 0.131 [115, 300] loss: 0.138 [115, 360] loss: 0.134 Epoch: 115 -> Loss: 0.206795409322 Epoch: 115 -> Test Accuracy: 90.715 [116, 60] loss: 0.112 [116, 120] loss: 0.129 [116, 180] loss: 0.134 [116, 240] loss: 0.130 [116, 300] loss: 0.139 [116, 360] loss: 0.135 Epoch: 116 -> Loss: 0.198866948485 Epoch: 116 -> Test Accuracy: 90.905 [117, 60] loss: 0.129 [117, 120] loss: 0.125 [117, 180] loss: 0.125 [117, 240] loss: 0.132 [117, 300] loss: 0.134 [117, 360] loss: 0.137 Epoch: 117 -> Loss: 0.152974322438 Epoch: 117 -> Test Accuracy: 90.8375 [118, 60] loss: 0.116 [118, 120] loss: 0.120 [118, 180] loss: 0.133 [118, 240] loss: 0.130 [118, 300] loss: 0.136 [118, 360] loss: 0.141 Epoch: 118 -> Loss: 0.0918317735195 Epoch: 118 -> Test Accuracy: 90.26 [119, 60] loss: 0.115 [119, 120] loss: 0.131 [119, 180] loss: 0.135 [119, 240] loss: 0.137 [119, 300] loss: 0.126 [119, 360] loss: 0.132 Epoch: 119 -> Loss: 0.214848846197 Epoch: 119 -> Test Accuracy: 90.455 [120, 60] loss: 0.124 [120, 120] loss: 0.127 [120, 180] loss: 0.131 [120, 240] loss: 0.135 [120, 300] loss: 0.136 [120, 360] loss: 0.139 Epoch: 120 -> Loss: 0.160332351923 Epoch: 120 -> Test Accuracy: 90.89 [121, 60] loss: 0.093 [121, 120] loss: 0.075 [121, 180] loss: 0.072 [121, 240] loss: 0.070 [121, 300] loss: 0.065 [121, 360] loss: 0.063 Epoch: 121 -> Loss: 0.0423624292016 Epoch: 121 -> Test Accuracy: 92.44 [122, 60] loss: 0.057 [122, 120] loss: 0.053 [122, 180] loss: 0.057 [122, 240] loss: 0.057 [122, 300] loss: 0.055 [122, 360] loss: 0.052 Epoch: 122 -> Loss: 0.0960680544376 Epoch: 122 -> Test Accuracy: 92.2625 [123, 60] loss: 0.046 [123, 120] loss: 0.046 [123, 180] loss: 0.050 [123, 240] loss: 0.050 [123, 300] loss: 0.048 [123, 360] loss: 0.048 Epoch: 123 -> Loss: 0.0571182966232 Epoch: 123 -> Test Accuracy: 92.36 [124, 60] loss: 0.043 [124, 120] loss: 0.044 [124, 180] loss: 0.043 [124, 240] loss: 0.045 [124, 300] loss: 0.045 [124, 360] loss: 0.044 Epoch: 124 -> Loss: 0.0507776737213 Epoch: 124 -> Test Accuracy: 92.2725 [125, 60] loss: 0.039 [125, 120] loss: 0.039 [125, 180] loss: 0.039 [125, 240] loss: 0.042 [125, 300] loss: 0.043 [125, 360] loss: 0.042 Epoch: 125 -> Loss: 0.0358629673719 Epoch: 125 -> Test Accuracy: 92.3075 [126, 60] loss: 0.042 [126, 120] loss: 0.038 [126, 180] loss: 0.039 [126, 240] loss: 0.040 [126, 300] loss: 0.038 [126, 360] loss: 0.035 Epoch: 126 -> Loss: 0.0723502635956 Epoch: 126 -> Test Accuracy: 92.44 [127, 60] loss: 0.033 [127, 120] loss: 0.032 [127, 180] loss: 0.036 [127, 240] loss: 0.036 [127, 300] loss: 0.038 [127, 360] loss: 0.040 Epoch: 127 -> Loss: 0.0361194983125 Epoch: 127 -> Test Accuracy: 92.2975 [128, 60] loss: 0.032 [128, 120] loss: 0.032 [128, 180] loss: 0.033 [128, 240] loss: 0.036 [128, 300] loss: 0.034 [128, 360] loss: 0.037 Epoch: 128 -> Loss: 0.0498210191727 Epoch: 128 -> Test Accuracy: 92.455 [129, 60] loss: 0.031 [129, 120] loss: 0.035 [129, 180] loss: 0.030 [129, 240] loss: 0.033 [129, 300] loss: 0.030 [129, 360] loss: 0.034 Epoch: 129 -> Loss: 0.0112496130168 Epoch: 129 -> Test Accuracy: 92.3725 [130, 60] loss: 0.029 [130, 120] loss: 0.032 [130, 180] loss: 0.033 [130, 240] loss: 0.030 [130, 300] loss: 0.033 [130, 360] loss: 0.031 Epoch: 130 -> Loss: 0.0396613068879 Epoch: 130 -> Test Accuracy: 92.2825 [131, 60] loss: 0.029 [131, 120] loss: 0.030 [131, 180] loss: 0.029 [131, 240] loss: 0.032 [131, 300] loss: 0.029 [131, 360] loss: 0.031 Epoch: 131 -> Loss: 0.0441687554121 Epoch: 131 -> Test Accuracy: 92.3325 [132, 60] loss: 0.029 [132, 120] loss: 0.028 [132, 180] loss: 0.029 [132, 240] loss: 0.030 [132, 300] loss: 0.033 [132, 360] loss: 0.030 Epoch: 132 -> Loss: 0.0172540359199 Epoch: 132 -> Test Accuracy: 92.155 [133, 60] loss: 0.028 [133, 120] loss: 0.029 [133, 180] loss: 0.030 [133, 240] loss: 0.031 [133, 300] loss: 0.031 [133, 360] loss: 0.028 Epoch: 133 -> Loss: 0.0438217520714 Epoch: 133 -> Test Accuracy: 92.26 [134, 60] loss: 0.027 [134, 120] loss: 0.027 [134, 180] loss: 0.028 [134, 240] loss: 0.027 [134, 300] loss: 0.030 [134, 360] loss: 0.030 Epoch: 134 -> Loss: 0.0326785072684 Epoch: 134 -> Test Accuracy: 92.4925 [135, 60] loss: 0.027 [135, 120] loss: 0.027 [135, 180] loss: 0.028 [135, 240] loss: 0.029 [135, 300] loss: 0.029 [135, 360] loss: 0.029 Epoch: 135 -> Loss: 0.0124627845362 Epoch: 135 -> Test Accuracy: 92.2775 [136, 60] loss: 0.023 [136, 120] loss: 0.027 [136, 180] loss: 0.026 [136, 240] loss: 0.027 [136, 300] loss: 0.029 [136, 360] loss: 0.031 Epoch: 136 -> Loss: 0.0220747478306 Epoch: 136 -> Test Accuracy: 92.11 [137, 60] loss: 0.022 [137, 120] loss: 0.025 [137, 180] loss: 0.028 [137, 240] loss: 0.024 [137, 300] loss: 0.029 [137, 360] loss: 0.028 Epoch: 137 -> Loss: 0.0135794626549 Epoch: 137 -> Test Accuracy: 92.16 [138, 60] loss: 0.025 [138, 120] loss: 0.027 [138, 180] loss: 0.026 [138, 240] loss: 0.026 [138, 300] loss: 0.025 [138, 360] loss: 0.028 Epoch: 138 -> Loss: 0.0104033928365 Epoch: 138 -> Test Accuracy: 92.1675 [139, 60] loss: 0.024 [139, 120] loss: 0.025 [139, 180] loss: 0.025 [139, 240] loss: 0.026 [139, 300] loss: 0.028 [139, 360] loss: 0.024 Epoch: 139 -> Loss: 0.0127109345049 Epoch: 139 -> Test Accuracy: 92.1775 [140, 60] loss: 0.021 [140, 120] loss: 0.021 [140, 180] loss: 0.026 [140, 240] loss: 0.026 [140, 300] loss: 0.025 [140, 360] loss: 0.026 Epoch: 140 -> Loss: 0.0257526282221 Epoch: 140 -> Test Accuracy: 92.115 [141, 60] loss: 0.023 [141, 120] loss: 0.023 [141, 180] loss: 0.026 [141, 240] loss: 0.026 [141, 300] loss: 0.024 [141, 360] loss: 0.027 Epoch: 141 -> Loss: 0.0222018398345 Epoch: 141 -> Test Accuracy: 92.065 [142, 60] loss: 0.024 [142, 120] loss: 0.025 [142, 180] loss: 0.026 [142, 240] loss: 0.024 [142, 300] loss: 0.025 [142, 360] loss: 0.024 Epoch: 142 -> Loss: 0.0381575599313 Epoch: 142 -> Test Accuracy: 92.0 [143, 60] loss: 0.022 [143, 120] loss: 0.026 [143, 180] loss: 0.025 [143, 240] loss: 0.027 [143, 300] loss: 0.024 [143, 360] loss: 0.026 Epoch: 143 -> Loss: 0.0457306429744 Epoch: 143 -> Test Accuracy: 92.1475 [144, 60] loss: 0.020 [144, 120] loss: 0.023 [144, 180] loss: 0.024 [144, 240] loss: 0.024 [144, 300] loss: 0.025 [144, 360] loss: 0.027 Epoch: 144 -> Loss: 0.0474306493998 Epoch: 144 -> Test Accuracy: 91.9325 [145, 60] loss: 0.022 [145, 120] loss: 0.023 [145, 180] loss: 0.023 [145, 240] loss: 0.024 [145, 300] loss: 0.024 [145, 360] loss: 0.026 Epoch: 145 -> Loss: 0.0229899715632 Epoch: 145 -> Test Accuracy: 92.1325 [146, 60] loss: 0.022 [146, 120] loss: 0.023 [146, 180] loss: 0.024 [146, 240] loss: 0.022 [146, 300] loss: 0.024 [146, 360] loss: 0.026 Epoch: 146 -> Loss: 0.0179903283715 Epoch: 146 -> Test Accuracy: 91.8925 [147, 60] loss: 0.024 [147, 120] loss: 0.025 [147, 180] loss: 0.023 [147, 240] loss: 0.024 [147, 300] loss: 0.026 [147, 360] loss: 0.025 Epoch: 147 -> Loss: 0.0645735636353 Epoch: 147 -> Test Accuracy: 91.975 [148, 60] loss: 0.023 [148, 120] loss: 0.022 [148, 180] loss: 0.023 [148, 240] loss: 0.026 [148, 300] loss: 0.023 [148, 360] loss: 0.026 Epoch: 148 -> Loss: 0.0240027327091 Epoch: 148 -> Test Accuracy: 91.765 [149, 60] loss: 0.023 [149, 120] loss: 0.021 [149, 180] loss: 0.022 [149, 240] loss: 0.025 [149, 300] loss: 0.026 [149, 360] loss: 0.026 Epoch: 149 -> Loss: 0.0485193766654 Epoch: 149 -> Test Accuracy: 92.1175 [150, 60] loss: 0.023 [150, 120] loss: 0.024 [150, 180] loss: 0.024 [150, 240] loss: 0.025 [150, 300] loss: 0.027 [150, 360] loss: 0.026 Epoch: 150 -> Loss: 0.0145528595895 Epoch: 150 -> Test Accuracy: 91.965 [151, 60] loss: 0.024 [151, 120] loss: 0.023 [151, 180] loss: 0.022 [151, 240] loss: 0.025 [151, 300] loss: 0.024 [151, 360] loss: 0.028 Epoch: 151 -> Loss: 0.0295480005443 Epoch: 151 -> Test Accuracy: 91.835 [152, 60] loss: 0.024 [152, 120] loss: 0.024 [152, 180] loss: 0.022 [152, 240] loss: 0.024 [152, 300] loss: 0.024 [152, 360] loss: 0.028 Epoch: 152 -> Loss: 0.0133929345757 Epoch: 152 -> Test Accuracy: 92.2025 [153, 60] loss: 0.024 [153, 120] loss: 0.025 [153, 180] loss: 0.026 [153, 240] loss: 0.025 [153, 300] loss: 0.027 [153, 360] loss: 0.028 Epoch: 153 -> Loss: 0.0198680497706 Epoch: 153 -> Test Accuracy: 91.95 [154, 60] loss: 0.024 [154, 120] loss: 0.024 [154, 180] loss: 0.025 [154, 240] loss: 0.024 [154, 300] loss: 0.027 [154, 360] loss: 0.026 Epoch: 154 -> Loss: 0.0222103130072 Epoch: 154 -> Test Accuracy: 91.965 [155, 60] loss: 0.022 [155, 120] loss: 0.022 [155, 180] loss: 0.024 [155, 240] loss: 0.025 [155, 300] loss: 0.029 [155, 360] loss: 0.028 Epoch: 155 -> Loss: 0.0241133067757 Epoch: 155 -> Test Accuracy: 91.935 [156, 60] loss: 0.024 [156, 120] loss: 0.026 [156, 180] loss: 0.024 [156, 240] loss: 0.026 [156, 300] loss: 0.024 [156, 360] loss: 0.024 Epoch: 156 -> Loss: 0.028401767835 Epoch: 156 -> Test Accuracy: 91.945 [157, 60] loss: 0.021 [157, 120] loss: 0.025 [157, 180] loss: 0.026 [157, 240] loss: 0.024 [157, 300] loss: 0.025 [157, 360] loss: 0.025 Epoch: 157 -> Loss: 0.0401775017381 Epoch: 157 -> Test Accuracy: 91.9725 [158, 60] loss: 0.022 [158, 120] loss: 0.023 [158, 180] loss: 0.025 [158, 240] loss: 0.025 [158, 300] loss: 0.026 [158, 360] loss: 0.024 Epoch: 158 -> Loss: 0.0298565886915 Epoch: 158 -> Test Accuracy: 91.9175 [159, 60] loss: 0.022 [159, 120] loss: 0.021 [159, 180] loss: 0.024 [159, 240] loss: 0.025 [159, 300] loss: 0.024 [159, 360] loss: 0.024 Epoch: 159 -> Loss: 0.0185794588178 Epoch: 159 -> Test Accuracy: 91.9525 [160, 60] loss: 0.024 [160, 120] loss: 0.022 [160, 180] loss: 0.023 [160, 240] loss: 0.025 [160, 300] loss: 0.026 [160, 360] loss: 0.026 Epoch: 160 -> Loss: 0.0450544841588 Epoch: 160 -> Test Accuracy: 91.885 [161, 60] loss: 0.019 [161, 120] loss: 0.016 [161, 180] loss: 0.016 [161, 240] loss: 0.014 [161, 300] loss: 0.014 [161, 360] loss: 0.014 Epoch: 161 -> Loss: 0.0197333395481 Epoch: 161 -> Test Accuracy: 92.43 [162, 60] loss: 0.012 [162, 120] loss: 0.012 [162, 180] loss: 0.013 [162, 240] loss: 0.012 [162, 300] loss: 0.012 [162, 360] loss: 0.011 Epoch: 162 -> Loss: 0.0110598672181 Epoch: 162 -> Test Accuracy: 92.4975 [163, 60] loss: 0.011 [163, 120] loss: 0.010 [163, 180] loss: 0.010 [163, 240] loss: 0.011 [163, 300] loss: 0.010 [163, 360] loss: 0.011 Epoch: 163 -> Loss: 0.00951013527811 Epoch: 163 -> Test Accuracy: 92.415 [164, 60] loss: 0.010 [164, 120] loss: 0.010 [164, 180] loss: 0.011 [164, 240] loss: 0.010 [164, 300] loss: 0.009 [164, 360] loss: 0.011 Epoch: 164 -> Loss: 0.0117461169139 Epoch: 164 -> Test Accuracy: 92.51 [165, 60] loss: 0.009 [165, 120] loss: 0.009 [165, 180] loss: 0.010 [165, 240] loss: 0.011 [165, 300] loss: 0.009 [165, 360] loss: 0.010 Epoch: 165 -> Loss: 0.00752267753705 Epoch: 165 -> Test Accuracy: 92.5775 [166, 60] loss: 0.009 [166, 120] loss: 0.008 [166, 180] loss: 0.010 [166, 240] loss: 0.010 [166, 300] loss: 0.009 [166, 360] loss: 0.008 Epoch: 166 -> Loss: 0.0189105384052 Epoch: 166 -> Test Accuracy: 92.48 [167, 60] loss: 0.009 [167, 120] loss: 0.009 [167, 180] loss: 0.008 [167, 240] loss: 0.009 [167, 300] loss: 0.009 [167, 360] loss: 0.007 Epoch: 167 -> Loss: 0.00751985330135 Epoch: 167 -> Test Accuracy: 92.585 [168, 60] loss: 0.008 [168, 120] loss: 0.008 [168, 180] loss: 0.008 [168, 240] loss: 0.009 [168, 300] loss: 0.008 [168, 360] loss: 0.009 Epoch: 168 -> Loss: 0.0115286121145 Epoch: 168 -> Test Accuracy: 92.5175 [169, 60] loss: 0.007 [169, 120] loss: 0.009 [169, 180] loss: 0.007 [169, 240] loss: 0.009 [169, 300] loss: 0.008 [169, 360] loss: 0.008 Epoch: 169 -> Loss: 0.00585957989097 Epoch: 169 -> Test Accuracy: 92.595 [170, 60] loss: 0.008 [170, 120] loss: 0.008 [170, 180] loss: 0.008 [170, 240] loss: 0.009 [170, 300] loss: 0.008 [170, 360] loss: 0.007 Epoch: 170 -> Loss: 0.0213364996016 Epoch: 170 -> Test Accuracy: 92.485 [171, 60] loss: 0.008 [171, 120] loss: 0.008 [171, 180] loss: 0.007 [171, 240] loss: 0.008 [171, 300] loss: 0.007 [171, 360] loss: 0.008 Epoch: 171 -> Loss: 0.00604963768274 Epoch: 171 -> Test Accuracy: 92.5125 [172, 60] loss: 0.007 [172, 120] loss: 0.007 [172, 180] loss: 0.008 [172, 240] loss: 0.008 [172, 300] loss: 0.007 [172, 360] loss: 0.008 Epoch: 172 -> Loss: 0.0104969134554 Epoch: 172 -> Test Accuracy: 92.4425 [173, 60] loss: 0.008 [173, 120] loss: 0.008 [173, 180] loss: 0.007 [173, 240] loss: 0.008 [173, 300] loss: 0.007 [173, 360] loss: 0.007 Epoch: 173 -> Loss: 0.00278425146826 Epoch: 173 -> Test Accuracy: 92.4725 [174, 60] loss: 0.006 [174, 120] loss: 0.007 [174, 180] loss: 0.006 [174, 240] loss: 0.007 [174, 300] loss: 0.007 [174, 360] loss: 0.007 Epoch: 174 -> Loss: 0.00542673934251 Epoch: 174 -> Test Accuracy: 92.4625 [175, 60] loss: 0.007 [175, 120] loss: 0.007 [175, 180] loss: 0.007 [175, 240] loss: 0.006 [175, 300] loss: 0.006 [175, 360] loss: 0.007 Epoch: 175 -> Loss: 0.00884590484202 Epoch: 175 -> Test Accuracy: 92.495 [176, 60] loss: 0.007 [176, 120] loss: 0.006 [176, 180] loss: 0.007 [176, 240] loss: 0.007 [176, 300] loss: 0.006 [176, 360] loss: 0.006 Epoch: 176 -> Loss: 0.0113233039156 Epoch: 176 -> Test Accuracy: 92.5075 [177, 60] loss: 0.007 [177, 120] loss: 0.006 [177, 180] loss: 0.006 [177, 240] loss: 0.006 [177, 300] loss: 0.007 [177, 360] loss: 0.008 Epoch: 177 -> Loss: 0.00249320571311 Epoch: 177 -> Test Accuracy: 92.5775 [178, 60] loss: 0.007 [178, 120] loss: 0.007 [178, 180] loss: 0.007 [178, 240] loss: 0.006 [178, 300] loss: 0.007 [178, 360] loss: 0.006 Epoch: 178 -> Loss: 0.00494795758277 Epoch: 178 -> Test Accuracy: 92.485 [179, 60] loss: 0.006 [179, 120] loss: 0.006 [179, 180] loss: 0.006 [179, 240] loss: 0.007 [179, 300] loss: 0.007 [179, 360] loss: 0.006 Epoch: 179 -> Loss: 0.00537769403309 Epoch: 179 -> Test Accuracy: 92.5 [180, 60] loss: 0.006 [180, 120] loss: 0.007 [180, 180] loss: 0.006 [180, 240] loss: 0.007 [180, 300] loss: 0.006 [180, 360] loss: 0.006 Epoch: 180 -> Loss: 0.00434305658564 Epoch: 180 -> Test Accuracy: 92.535 [181, 60] loss: 0.006 [181, 120] loss: 0.007 [181, 180] loss: 0.006 [181, 240] loss: 0.007 [181, 300] loss: 0.007 [181, 360] loss: 0.006 Epoch: 181 -> Loss: 0.00808005221188 Epoch: 181 -> Test Accuracy: 92.5275 [182, 60] loss: 0.006 [182, 120] loss: 0.006 [182, 180] loss: 0.006 [182, 240] loss: 0.006 [182, 300] loss: 0.006 [182, 360] loss: 0.006 Epoch: 182 -> Loss: 0.00897836498916 Epoch: 182 -> Test Accuracy: 92.6 [183, 60] loss: 0.006 [183, 120] loss: 0.006 [183, 180] loss: 0.006 [183, 240] loss: 0.006 [183, 300] loss: 0.006 [183, 360] loss: 0.007 Epoch: 183 -> Loss: 0.00255749886855 Epoch: 183 -> Test Accuracy: 92.47 [184, 60] loss: 0.006 [184, 120] loss: 0.006 [184, 180] loss: 0.006 [184, 240] loss: 0.006 [184, 300] loss: 0.006 [184, 360] loss: 0.006 Epoch: 184 -> Loss: 0.00525060575455 Epoch: 184 -> Test Accuracy: 92.475 [185, 60] loss: 0.006 [185, 120] loss: 0.006 [185, 180] loss: 0.006 [185, 240] loss: 0.006 [185, 300] loss: 0.007 [185, 360] loss: 0.007 Epoch: 185 -> Loss: 0.0120564447716 Epoch: 185 -> Test Accuracy: 92.4775 [186, 60] loss: 0.005 [186, 120] loss: 0.006 [186, 180] loss: 0.006 [186, 240] loss: 0.006 [186, 300] loss: 0.006 [186, 360] loss: 0.006 Epoch: 186 -> Loss: 0.00262727355585 Epoch: 186 -> Test Accuracy: 92.445 [187, 60] loss: 0.006 [187, 120] loss: 0.006 [187, 180] loss: 0.005 [187, 240] loss: 0.006 [187, 300] loss: 0.006 [187, 360] loss: 0.006 Epoch: 187 -> Loss: 0.00634869094938 Epoch: 187 -> Test Accuracy: 92.42 [188, 60] loss: 0.006 [188, 120] loss: 0.006 [188, 180] loss: 0.005 [188, 240] loss: 0.005 [188, 300] loss: 0.006 [188, 360] loss: 0.006 Epoch: 188 -> Loss: 0.0084777334705 Epoch: 188 -> Test Accuracy: 92.455 [189, 60] loss: 0.006 [189, 120] loss: 0.005 [189, 180] loss: 0.006 [189, 240] loss: 0.006 [189, 300] loss: 0.006 [189, 360] loss: 0.005 Epoch: 189 -> Loss: 0.00381430843845 Epoch: 189 -> Test Accuracy: 92.47 [190, 60] loss: 0.006 [190, 120] loss: 0.006 [190, 180] loss: 0.005 [190, 240] loss: 0.005 [190, 300] loss: 0.006 [190, 360] loss: 0.005 Epoch: 190 -> Loss: 0.00268593197688 Epoch: 190 -> Test Accuracy: 92.51 [191, 60] loss: 0.006 [191, 120] loss: 0.006 [191, 180] loss: 0.006 [191, 240] loss: 0.005 [191, 300] loss: 0.005 [191, 360] loss: 0.006 Epoch: 191 -> Loss: 0.0037877925206 Epoch: 191 -> Test Accuracy: 92.545 [192, 60] loss: 0.005 [192, 120] loss: 0.005 [192, 180] loss: 0.006 [192, 240] loss: 0.005 [192, 300] loss: 0.005 [192, 360] loss: 0.006 Epoch: 192 -> Loss: 0.0051264828071 Epoch: 192 -> Test Accuracy: 92.4 [193, 60] loss: 0.005 [193, 120] loss: 0.005 [193, 180] loss: 0.006 [193, 240] loss: 0.006 [193, 300] loss: 0.006 [193, 360] loss: 0.006 Epoch: 193 -> Loss: 0.00503174727783 Epoch: 193 -> Test Accuracy: 92.445 [194, 60] loss: 0.005 [194, 120] loss: 0.005 [194, 180] loss: 0.005 [194, 240] loss: 0.006 [194, 300] loss: 0.006 [194, 360] loss: 0.005 Epoch: 194 -> Loss: 0.00384769542143 Epoch: 194 -> Test Accuracy: 92.54 [195, 60] loss: 0.006 [195, 120] loss: 0.005 [195, 180] loss: 0.006 [195, 240] loss: 0.006 [195, 300] loss: 0.005 [195, 360] loss: 0.005 Epoch: 195 -> Loss: 0.00383017142303 Epoch: 195 -> Test Accuracy: 92.5175 [196, 60] loss: 0.005 [196, 120] loss: 0.005 [196, 180] loss: 0.006 [196, 240] loss: 0.005 [196, 300] loss: 0.005 [196, 360] loss: 0.005 Epoch: 196 -> Loss: 0.0110269840807 Epoch: 196 -> Test Accuracy: 92.535 [197, 60] loss: 0.005 [197, 120] loss: 0.005 [197, 180] loss: 0.006 [197, 240] loss: 0.005 [197, 300] loss: 0.006 [197, 360] loss: 0.005 Epoch: 197 -> Loss: 0.00754360621795 Epoch: 197 -> Test Accuracy: 92.5475 [198, 60] loss: 0.005 [198, 120] loss: 0.006 [198, 180] loss: 0.005 [198, 240] loss: 0.005 [198, 300] loss: 0.006 [198, 360] loss: 0.005 Epoch: 198 -> Loss: 0.0103376815096 Epoch: 198 -> Test Accuracy: 92.6025 [199, 60] loss: 0.005 [199, 120] loss: 0.005 [199, 180] loss: 0.005 [199, 240] loss: 0.005 [199, 300] loss: 0.006 [199, 360] loss: 0.006 Epoch: 199 -> Loss: 0.00404023518786 Epoch: 199 -> Test Accuracy: 92.64 [200, 60] loss: 0.006 [200, 120] loss: 0.005 [200, 180] loss: 0.005 [200, 240] loss: 0.005 [200, 300] loss: 0.005 [200, 360] loss: 0.005 Epoch: 200 -> Loss: 0.00244236225262 Epoch: 200 -> Test Accuracy: 92.6325 Finished Training
In [8]:
# train NonLinearClassifiers on feature map of net_3block
block4_loss_log, _, block4_test_accuracy_log, _, _ = tr.train_all_blocks(4, 10, [0.1, 0.02, 0.004, 0.0008],
[20, 40, 45, 100], 0.9, 5e-4, net_block4, criterion, trainloader, None, testloader)
[1, 60] loss: 2.230 [1, 120] loss: 1.261 [1, 180] loss: 1.156 [1, 240] loss: 1.095 [1, 300] loss: 1.038 [1, 360] loss: 1.018 Epoch: 1 -> Loss: 0.952750682831 Epoch: 1 -> Test Accuracy: 67.22 [2, 60] loss: 0.948 [2, 120] loss: 0.923 [2, 180] loss: 0.898 [2, 240] loss: 0.926 [2, 300] loss: 0.883 [2, 360] loss: 0.866 Epoch: 2 -> Loss: 0.915363311768 Epoch: 2 -> Test Accuracy: 70.55 [3, 60] loss: 0.825 [3, 120] loss: 0.847 [3, 180] loss: 0.832 [3, 240] loss: 0.823 [3, 300] loss: 0.791 [3, 360] loss: 0.810 Epoch: 3 -> Loss: 0.695132553577 Epoch: 3 -> Test Accuracy: 73.29 [4, 60] loss: 0.773 [4, 120] loss: 0.784 [4, 180] loss: 0.756 [4, 240] loss: 0.784 [4, 300] loss: 0.762 [4, 360] loss: 0.753 Epoch: 4 -> Loss: 0.705285429955 Epoch: 4 -> Test Accuracy: 73.98 [5, 60] loss: 0.727 [5, 120] loss: 0.750 [5, 180] loss: 0.763 [5, 240] loss: 0.721 [5, 300] loss: 0.723 [5, 360] loss: 0.717 Epoch: 5 -> Loss: 0.739547371864 Epoch: 5 -> Test Accuracy: 75.62 [6, 60] loss: 0.721 [6, 120] loss: 0.724 [6, 180] loss: 0.711 [6, 240] loss: 0.686 [6, 300] loss: 0.712 [6, 360] loss: 0.714 Epoch: 6 -> Loss: 0.644628822803 Epoch: 6 -> Test Accuracy: 76.28 [7, 60] loss: 0.690 [7, 120] loss: 0.687 [7, 180] loss: 0.693 [7, 240] loss: 0.694 [7, 300] loss: 0.687 [7, 360] loss: 0.693 Epoch: 7 -> Loss: 0.792090713978 Epoch: 7 -> Test Accuracy: 76.84 [8, 60] loss: 0.661 [8, 120] loss: 0.686 [8, 180] loss: 0.675 [8, 240] loss: 0.682 [8, 300] loss: 0.680 [8, 360] loss: 0.691 Epoch: 8 -> Loss: 0.655223548412 Epoch: 8 -> Test Accuracy: 77.01 [9, 60] loss: 0.650 [9, 120] loss: 0.663 [9, 180] loss: 0.669 [9, 240] loss: 0.655 [9, 300] loss: 0.657 [9, 360] loss: 0.676 Epoch: 9 -> Loss: 0.560119509697 Epoch: 9 -> Test Accuracy: 77.28 [10, 60] loss: 0.642 [10, 120] loss: 0.629 [10, 180] loss: 0.654 [10, 240] loss: 0.669 [10, 300] loss: 0.662 [10, 360] loss: 0.669 Epoch: 10 -> Loss: 0.82323038578 Epoch: 10 -> Test Accuracy: 77.14 [11, 60] loss: 0.643 [11, 120] loss: 0.632 [11, 180] loss: 0.653 [11, 240] loss: 0.662 [11, 300] loss: 0.641 [11, 360] loss: 0.632 Epoch: 11 -> Loss: 0.432897984982 Epoch: 11 -> Test Accuracy: 78.05 [12, 60] loss: 0.632 [12, 120] loss: 0.654 [12, 180] loss: 0.639 [12, 240] loss: 0.633 [12, 300] loss: 0.639 [12, 360] loss: 0.641 Epoch: 12 -> Loss: 0.653931498528 Epoch: 12 -> Test Accuracy: 78.27 [13, 60] loss: 0.603 [13, 120] loss: 0.627 [13, 180] loss: 0.619 [13, 240] loss: 0.654 [13, 300] loss: 0.644 [13, 360] loss: 0.640 Epoch: 13 -> Loss: 0.710495054722 Epoch: 13 -> Test Accuracy: 78.23 [14, 60] loss: 0.633 [14, 120] loss: 0.618 [14, 180] loss: 0.613 [14, 240] loss: 0.637 [14, 300] loss: 0.633 [14, 360] loss: 0.656 Epoch: 14 -> Loss: 0.647937297821 Epoch: 14 -> Test Accuracy: 78.51 [15, 60] loss: 0.613 [15, 120] loss: 0.609 [15, 180] loss: 0.621 [15, 240] loss: 0.637 [15, 300] loss: 0.627 [15, 360] loss: 0.629 Epoch: 15 -> Loss: 0.76855301857 Epoch: 15 -> Test Accuracy: 78.02 [16, 60] loss: 0.606 [16, 120] loss: 0.629 [16, 180] loss: 0.621 [16, 240] loss: 0.595 [16, 300] loss: 0.620 [16, 360] loss: 0.647 Epoch: 16 -> Loss: 0.796156644821 Epoch: 16 -> Test Accuracy: 78.29 [17, 60] loss: 0.602 [17, 120] loss: 0.599 [17, 180] loss: 0.622 [17, 240] loss: 0.638 [17, 300] loss: 0.627 [17, 360] loss: 0.620 Epoch: 17 -> Loss: 0.669784069061 Epoch: 17 -> Test Accuracy: 78.06 [18, 60] loss: 0.598 [18, 120] loss: 0.610 [18, 180] loss: 0.604 [18, 240] loss: 0.603 [18, 300] loss: 0.635 [18, 360] loss: 0.634 Epoch: 18 -> Loss: 0.724856078625 Epoch: 18 -> Test Accuracy: 78.13 [19, 60] loss: 0.588 [19, 120] loss: 0.603 [19, 180] loss: 0.626 [19, 240] loss: 0.625 [19, 300] loss: 0.598 [19, 360] loss: 0.619 Epoch: 19 -> Loss: 0.865086734295 Epoch: 19 -> Test Accuracy: 78.23 [20, 60] loss: 0.592 [20, 120] loss: 0.606 [20, 180] loss: 0.617 [20, 240] loss: 0.602 [20, 300] loss: 0.616 [20, 360] loss: 0.637 Epoch: 20 -> Loss: 0.652887940407 Epoch: 20 -> Test Accuracy: 79.02 [21, 60] loss: 0.556 [21, 120] loss: 0.510 [21, 180] loss: 0.510 [21, 240] loss: 0.502 [21, 300] loss: 0.501 [21, 360] loss: 0.504 Epoch: 21 -> Loss: 0.482809841633 Epoch: 21 -> Test Accuracy: 81.46 [22, 60] loss: 0.464 [22, 120] loss: 0.485 [22, 180] loss: 0.473 [22, 240] loss: 0.482 [22, 300] loss: 0.465 [22, 360] loss: 0.486 Epoch: 22 -> Loss: 0.485416263342 Epoch: 22 -> Test Accuracy: 81.38 [23, 60] loss: 0.441 [23, 120] loss: 0.456 [23, 180] loss: 0.465 [23, 240] loss: 0.470 [23, 300] loss: 0.446 [23, 360] loss: 0.464 Epoch: 23 -> Loss: 0.383998095989 Epoch: 23 -> Test Accuracy: 81.89 [24, 60] loss: 0.440 [24, 120] loss: 0.444 [24, 180] loss: 0.447 [24, 240] loss: 0.456 [24, 300] loss: 0.459 [24, 360] loss: 0.452 Epoch: 24 -> Loss: 0.478100955486 Epoch: 24 -> Test Accuracy: 81.64 [25, 60] loss: 0.444 [25, 120] loss: 0.430 [25, 180] loss: 0.439 [25, 240] loss: 0.432 [25, 300] loss: 0.448 [25, 360] loss: 0.445 Epoch: 25 -> Loss: 0.58848541975 Epoch: 25 -> Test Accuracy: 81.59 [26, 60] loss: 0.420 [26, 120] loss: 0.424 [26, 180] loss: 0.440 [26, 240] loss: 0.433 [26, 300] loss: 0.430 [26, 360] loss: 0.436 Epoch: 26 -> Loss: 0.395174086094 Epoch: 26 -> Test Accuracy: 81.59 [27, 60] loss: 0.425 [27, 120] loss: 0.422 [27, 180] loss: 0.410 [27, 240] loss: 0.418 [27, 300] loss: 0.420 [27, 360] loss: 0.424 Epoch: 27 -> Loss: 0.599011421204 Epoch: 27 -> Test Accuracy: 81.62 [28, 60] loss: 0.413 [28, 120] loss: 0.419 [28, 180] loss: 0.413 [28, 240] loss: 0.423 [28, 300] loss: 0.435 [28, 360] loss: 0.429 Epoch: 28 -> Loss: 0.480876982212 Epoch: 28 -> Test Accuracy: 81.92 [29, 60] loss: 0.404 [29, 120] loss: 0.402 [29, 180] loss: 0.425 [29, 240] loss: 0.409 [29, 300] loss: 0.420 [29, 360] loss: 0.434 Epoch: 29 -> Loss: 0.37843477726 Epoch: 29 -> Test Accuracy: 81.7 [30, 60] loss: 0.421 [30, 120] loss: 0.407 [30, 180] loss: 0.405 [30, 240] loss: 0.399 [30, 300] loss: 0.424 [30, 360] loss: 0.415 Epoch: 30 -> Loss: 0.355725944042 Epoch: 30 -> Test Accuracy: 81.94 [31, 60] loss: 0.402 [31, 120] loss: 0.411 [31, 180] loss: 0.425 [31, 240] loss: 0.420 [31, 300] loss: 0.423 [31, 360] loss: 0.416 Epoch: 31 -> Loss: 0.418404638767 Epoch: 31 -> Test Accuracy: 82.11 [32, 60] loss: 0.422 [32, 120] loss: 0.402 [32, 180] loss: 0.410 [32, 240] loss: 0.397 [32, 300] loss: 0.415 [32, 360] loss: 0.425 Epoch: 32 -> Loss: 0.333952903748 Epoch: 32 -> Test Accuracy: 82.3 [33, 60] loss: 0.388 [33, 120] loss: 0.405 [33, 180] loss: 0.405 [33, 240] loss: 0.416 [33, 300] loss: 0.418 [33, 360] loss: 0.425 Epoch: 33 -> Loss: 0.731445670128 Epoch: 33 -> Test Accuracy: 81.97 [34, 60] loss: 0.397 [34, 120] loss: 0.385 [34, 180] loss: 0.425 [34, 240] loss: 0.406 [34, 300] loss: 0.407 [34, 360] loss: 0.415 Epoch: 34 -> Loss: 0.553569555283 Epoch: 34 -> Test Accuracy: 81.59 [35, 60] loss: 0.397 [35, 120] loss: 0.395 [35, 180] loss: 0.407 [35, 240] loss: 0.398 [35, 300] loss: 0.418 [35, 360] loss: 0.427 Epoch: 35 -> Loss: 0.419232696295 Epoch: 35 -> Test Accuracy: 81.89 [36, 60] loss: 0.415 [36, 120] loss: 0.402 [36, 180] loss: 0.385 [36, 240] loss: 0.387 [36, 300] loss: 0.397 [36, 360] loss: 0.397 Epoch: 36 -> Loss: 0.358517944813 Epoch: 36 -> Test Accuracy: 81.42 [37, 60] loss: 0.407 [37, 120] loss: 0.389 [37, 180] loss: 0.388 [37, 240] loss: 0.413 [37, 300] loss: 0.403 [37, 360] loss: 0.404 Epoch: 37 -> Loss: 0.481689363718 Epoch: 37 -> Test Accuracy: 81.52 [38, 60] loss: 0.390 [38, 120] loss: 0.404 [38, 180] loss: 0.417 [38, 240] loss: 0.406 [38, 300] loss: 0.397 [38, 360] loss: 0.419 Epoch: 38 -> Loss: 0.464514911175 Epoch: 38 -> Test Accuracy: 81.76 [39, 60] loss: 0.407 [39, 120] loss: 0.400 [39, 180] loss: 0.391 [39, 240] loss: 0.392 [39, 300] loss: 0.401 [39, 360] loss: 0.418 Epoch: 39 -> Loss: 0.431790441275 Epoch: 39 -> Test Accuracy: 81.0 [40, 60] loss: 0.380 [40, 120] loss: 0.392 [40, 180] loss: 0.404 [40, 240] loss: 0.398 [40, 300] loss: 0.424 [40, 360] loss: 0.412 Epoch: 40 -> Loss: 0.348495423794 Epoch: 40 -> Test Accuracy: 81.21 [41, 60] loss: 0.367 [41, 120] loss: 0.362 [41, 180] loss: 0.366 [41, 240] loss: 0.346 [41, 300] loss: 0.357 [41, 360] loss: 0.344 Epoch: 41 -> Loss: 0.338767826557 Epoch: 41 -> Test Accuracy: 82.69 [42, 60] loss: 0.332 [42, 120] loss: 0.328 [42, 180] loss: 0.324 [42, 240] loss: 0.348 [42, 300] loss: 0.337 [42, 360] loss: 0.324 Epoch: 42 -> Loss: 0.255785286427 Epoch: 42 -> Test Accuracy: 82.87 [43, 60] loss: 0.332 [43, 120] loss: 0.317 [43, 180] loss: 0.317 [43, 240] loss: 0.316 [43, 300] loss: 0.325 [43, 360] loss: 0.323 Epoch: 43 -> Loss: 0.271632134914 Epoch: 43 -> Test Accuracy: 82.77 [44, 60] loss: 0.330 [44, 120] loss: 0.324 [44, 180] loss: 0.328 [44, 240] loss: 0.304 [44, 300] loss: 0.319 [44, 360] loss: 0.306 Epoch: 44 -> Loss: 0.272089123726 Epoch: 44 -> Test Accuracy: 82.87 [45, 60] loss: 0.310 [45, 120] loss: 0.312 [45, 180] loss: 0.300 [45, 240] loss: 0.311 [45, 300] loss: 0.306 [45, 360] loss: 0.311 Epoch: 45 -> Loss: 0.231358855963 Epoch: 45 -> Test Accuracy: 82.81 [46, 60] loss: 0.289 [46, 120] loss: 0.314 [46, 180] loss: 0.315 [46, 240] loss: 0.302 [46, 300] loss: 0.289 [46, 360] loss: 0.301 Epoch: 46 -> Loss: 0.206209033728 Epoch: 46 -> Test Accuracy: 82.97 [47, 60] loss: 0.293 [47, 120] loss: 0.298 [47, 180] loss: 0.289 [47, 240] loss: 0.302 [47, 300] loss: 0.288 [47, 360] loss: 0.296 Epoch: 47 -> Loss: 0.462675005198 Epoch: 47 -> Test Accuracy: 82.83 [48, 60] loss: 0.293 [48, 120] loss: 0.296 [48, 180] loss: 0.289 [48, 240] loss: 0.298 [48, 300] loss: 0.287 [48, 360] loss: 0.283 Epoch: 48 -> Loss: 0.299165278673 Epoch: 48 -> Test Accuracy: 82.99 [49, 60] loss: 0.274 [49, 120] loss: 0.292 [49, 180] loss: 0.288 [49, 240] loss: 0.293 [49, 300] loss: 0.287 [49, 360] loss: 0.297 Epoch: 49 -> Loss: 0.173502907157 Epoch: 49 -> Test Accuracy: 83.16 [50, 60] loss: 0.288 [50, 120] loss: 0.282 [50, 180] loss: 0.273 [50, 240] loss: 0.290 [50, 300] loss: 0.294 [50, 360] loss: 0.286 Epoch: 50 -> Loss: 0.494676679373 Epoch: 50 -> Test Accuracy: 83.08 [51, 60] loss: 0.290 [51, 120] loss: 0.284 [51, 180] loss: 0.283 [51, 240] loss: 0.290 [51, 300] loss: 0.293 [51, 360] loss: 0.275 Epoch: 51 -> Loss: 0.411444842815 Epoch: 51 -> Test Accuracy: 83.12 [52, 60] loss: 0.284 [52, 120] loss: 0.278 [52, 180] loss: 0.300 [52, 240] loss: 0.293 [52, 300] loss: 0.270 [52, 360] loss: 0.276 Epoch: 52 -> Loss: 0.199544996023 Epoch: 52 -> Test Accuracy: 83.23 [53, 60] loss: 0.287 [53, 120] loss: 0.280 [53, 180] loss: 0.289 [53, 240] loss: 0.288 [53, 300] loss: 0.285 [53, 360] loss: 0.290 Epoch: 53 -> Loss: 0.387977451086 Epoch: 53 -> Test Accuracy: 83.07 [54, 60] loss: 0.263 [54, 120] loss: 0.291 [54, 180] loss: 0.286 [54, 240] loss: 0.273 [54, 300] loss: 0.306 [54, 360] loss: 0.282 Epoch: 54 -> Loss: 0.397488355637 Epoch: 54 -> Test Accuracy: 83.16 [55, 60] loss: 0.292 [55, 120] loss: 0.288 [55, 180] loss: 0.276 [55, 240] loss: 0.265 [55, 300] loss: 0.274 [55, 360] loss: 0.267 Epoch: 55 -> Loss: 0.258340984583 Epoch: 55 -> Test Accuracy: 83.17 [56, 60] loss: 0.284 [56, 120] loss: 0.268 [56, 180] loss: 0.280 [56, 240] loss: 0.282 [56, 300] loss: 0.279 [56, 360] loss: 0.272 Epoch: 56 -> Loss: 0.32501745224 Epoch: 56 -> Test Accuracy: 83.29 [57, 60] loss: 0.263 [57, 120] loss: 0.275 [57, 180] loss: 0.272 [57, 240] loss: 0.279 [57, 300] loss: 0.291 [57, 360] loss: 0.279 Epoch: 57 -> Loss: 0.295313954353 Epoch: 57 -> Test Accuracy: 83.35 [58, 60] loss: 0.276 [58, 120] loss: 0.277 [58, 180] loss: 0.272 [58, 240] loss: 0.259 [58, 300] loss: 0.274 [58, 360] loss: 0.267 Epoch: 58 -> Loss: 0.275536715984 Epoch: 58 -> Test Accuracy: 83.25 [59, 60] loss: 0.272 [59, 120] loss: 0.274 [59, 180] loss: 0.291 [59, 240] loss: 0.280 [59, 300] loss: 0.272 [59, 360] loss: 0.275 Epoch: 59 -> Loss: 0.297197401524 Epoch: 59 -> Test Accuracy: 83.2 [60, 60] loss: 0.273 [60, 120] loss: 0.280 [60, 180] loss: 0.284 [60, 240] loss: 0.279 [60, 300] loss: 0.277 [60, 360] loss: 0.258 Epoch: 60 -> Loss: 0.14818200469 Epoch: 60 -> Test Accuracy: 83.14 [61, 60] loss: 0.277 [61, 120] loss: 0.273 [61, 180] loss: 0.280 [61, 240] loss: 0.281 [61, 300] loss: 0.272 [61, 360] loss: 0.275 Epoch: 61 -> Loss: 0.30923396349 Epoch: 61 -> Test Accuracy: 83.18 [62, 60] loss: 0.266 [62, 120] loss: 0.260 [62, 180] loss: 0.271 [62, 240] loss: 0.273 [62, 300] loss: 0.270 [62, 360] loss: 0.291 Epoch: 62 -> Loss: 0.328295975924 Epoch: 62 -> Test Accuracy: 83.0 [63, 60] loss: 0.264 [63, 120] loss: 0.271 [63, 180] loss: 0.275 [63, 240] loss: 0.277 [63, 300] loss: 0.266 [63, 360] loss: 0.269 Epoch: 63 -> Loss: 0.30826947093 Epoch: 63 -> Test Accuracy: 83.21 [64, 60] loss: 0.266 [64, 120] loss: 0.265 [64, 180] loss: 0.292 [64, 240] loss: 0.264 [64, 300] loss: 0.276 [64, 360] loss: 0.274 Epoch: 64 -> Loss: 0.30528062582 Epoch: 64 -> Test Accuracy: 83.03 [65, 60] loss: 0.284 [65, 120] loss: 0.262 [65, 180] loss: 0.264 [65, 240] loss: 0.270 [65, 300] loss: 0.266 [65, 360] loss: 0.274 Epoch: 65 -> Loss: 0.249018624425 Epoch: 65 -> Test Accuracy: 83.16 [66, 60] loss: 0.268 [66, 120] loss: 0.277 [66, 180] loss: 0.259 [66, 240] loss: 0.281 [66, 300] loss: 0.261 [66, 360] loss: 0.274 Epoch: 66 -> Loss: 0.274273097515 Epoch: 66 -> Test Accuracy: 83.16 [67, 60] loss: 0.263 [67, 120] loss: 0.260 [67, 180] loss: 0.280 [67, 240] loss: 0.273 [67, 300] loss: 0.256 [67, 360] loss: 0.267 Epoch: 67 -> Loss: 0.17559632659 Epoch: 67 -> Test Accuracy: 83.24 [68, 60] loss: 0.265 [68, 120] loss: 0.260 [68, 180] loss: 0.259 [68, 240] loss: 0.273 [68, 300] loss: 0.271 [68, 360] loss: 0.274 Epoch: 68 -> Loss: 0.307023853064 Epoch: 68 -> Test Accuracy: 83.29 [69, 60] loss: 0.258 [69, 120] loss: 0.261 [69, 180] loss: 0.264 [69, 240] loss: 0.265 [69, 300] loss: 0.278 [69, 360] loss: 0.262 Epoch: 69 -> Loss: 0.280606687069 Epoch: 69 -> Test Accuracy: 83.34 [70, 60] loss: 0.264 [70, 120] loss: 0.263 [70, 180] loss: 0.256 [70, 240] loss: 0.269 [70, 300] loss: 0.256 [70, 360] loss: 0.275 Epoch: 70 -> Loss: 0.251718312502 Epoch: 70 -> Test Accuracy: 83.33 [71, 60] loss: 0.261 [71, 120] loss: 0.259 [71, 180] loss: 0.263 [71, 240] loss: 0.270 [71, 300] loss: 0.266 [71, 360] loss: 0.257 Epoch: 71 -> Loss: 0.174997925758 Epoch: 71 -> Test Accuracy: 83.24 [72, 60] loss: 0.256 [72, 120] loss: 0.257 [72, 180] loss: 0.255 [72, 240] loss: 0.268 [72, 300] loss: 0.265 [72, 360] loss: 0.253 Epoch: 72 -> Loss: 0.235711380839 Epoch: 72 -> Test Accuracy: 83.09 [73, 60] loss: 0.254 [73, 120] loss: 0.255 [73, 180] loss: 0.264 [73, 240] loss: 0.264 [73, 300] loss: 0.277 [73, 360] loss: 0.262 Epoch: 73 -> Loss: 0.318305492401 Epoch: 73 -> Test Accuracy: 83.21 [74, 60] loss: 0.266 [74, 120] loss: 0.255 [74, 180] loss: 0.256 [74, 240] loss: 0.262 [74, 300] loss: 0.266 [74, 360] loss: 0.258 Epoch: 74 -> Loss: 0.349063485861 Epoch: 74 -> Test Accuracy: 83.21 [75, 60] loss: 0.247 [75, 120] loss: 0.255 [75, 180] loss: 0.253 [75, 240] loss: 0.254 [75, 300] loss: 0.269 [75, 360] loss: 0.265 Epoch: 75 -> Loss: 0.211874723434 Epoch: 75 -> Test Accuracy: 83.07 [76, 60] loss: 0.252 [76, 120] loss: 0.263 [76, 180] loss: 0.258 [76, 240] loss: 0.257 [76, 300] loss: 0.267 [76, 360] loss: 0.261 Epoch: 76 -> Loss: 0.393870055676 Epoch: 76 -> Test Accuracy: 83.02 [77, 60] loss: 0.258 [77, 120] loss: 0.256 [77, 180] loss: 0.259 [77, 240] loss: 0.253 [77, 300] loss: 0.270 [77, 360] loss: 0.255 Epoch: 77 -> Loss: 0.254414737225 Epoch: 77 -> Test Accuracy: 83.17 [78, 60] loss: 0.248 [78, 120] loss: 0.264 [78, 180] loss: 0.256 [78, 240] loss: 0.251 [78, 300] loss: 0.262 [78, 360] loss: 0.249 Epoch: 78 -> Loss: 0.285253852606 Epoch: 78 -> Test Accuracy: 83.13 [79, 60] loss: 0.254 [79, 120] loss: 0.248 [79, 180] loss: 0.262 [79, 240] loss: 0.267 [79, 300] loss: 0.248 [79, 360] loss: 0.248 Epoch: 79 -> Loss: 0.229285866022 Epoch: 79 -> Test Accuracy: 83.03 [80, 60] loss: 0.272 [80, 120] loss: 0.261 [80, 180] loss: 0.259 [80, 240] loss: 0.258 [80, 300] loss: 0.254 [80, 360] loss: 0.249 Epoch: 80 -> Loss: 0.279805004597 Epoch: 80 -> Test Accuracy: 82.96 [81, 60] loss: 0.251 [81, 120] loss: 0.251 [81, 180] loss: 0.250 [81, 240] loss: 0.256 [81, 300] loss: 0.259 [81, 360] loss: 0.247 Epoch: 81 -> Loss: 0.243606477976 Epoch: 81 -> Test Accuracy: 82.93 [82, 60] loss: 0.249 [82, 120] loss: 0.248 [82, 180] loss: 0.254 [82, 240] loss: 0.252 [82, 300] loss: 0.249 [82, 360] loss: 0.251 Epoch: 82 -> Loss: 0.286044418812 Epoch: 82 -> Test Accuracy: 83.18 [83, 60] loss: 0.256 [83, 120] loss: 0.253 [83, 180] loss: 0.258 [83, 240] loss: 0.255 [83, 300] loss: 0.249 [83, 360] loss: 0.242 Epoch: 83 -> Loss: 0.435457646847 Epoch: 83 -> Test Accuracy: 83.14 [84, 60] loss: 0.259 [84, 120] loss: 0.260 [84, 180] loss: 0.248 [84, 240] loss: 0.250 [84, 300] loss: 0.261 [84, 360] loss: 0.249 Epoch: 84 -> Loss: 0.239048808813 Epoch: 84 -> Test Accuracy: 83.15 [85, 60] loss: 0.262 [85, 120] loss: 0.244 [85, 180] loss: 0.252 [85, 240] loss: 0.256 [85, 300] loss: 0.252 [85, 360] loss: 0.245 Epoch: 85 -> Loss: 0.258603096008 Epoch: 85 -> Test Accuracy: 83.1 [86, 60] loss: 0.247 [86, 120] loss: 0.259 [86, 180] loss: 0.251 [86, 240] loss: 0.243 [86, 300] loss: 0.245 [86, 360] loss: 0.257 Epoch: 86 -> Loss: 0.231459587812 Epoch: 86 -> Test Accuracy: 82.98 [87, 60] loss: 0.246 [87, 120] loss: 0.248 [87, 180] loss: 0.243 [87, 240] loss: 0.247 [87, 300] loss: 0.259 [87, 360] loss: 0.238 Epoch: 87 -> Loss: 0.371732950211 Epoch: 87 -> Test Accuracy: 83.05 [88, 60] loss: 0.241 [88, 120] loss: 0.249 [88, 180] loss: 0.246 [88, 240] loss: 0.243 [88, 300] loss: 0.248 [88, 360] loss: 0.245 Epoch: 88 -> Loss: 0.245003938675 Epoch: 88 -> Test Accuracy: 83.12 [89, 60] loss: 0.244 [89, 120] loss: 0.239 [89, 180] loss: 0.256 [89, 240] loss: 0.243 [89, 300] loss: 0.239 [89, 360] loss: 0.242 Epoch: 89 -> Loss: 0.161676540971 Epoch: 89 -> Test Accuracy: 82.96 [90, 60] loss: 0.242 [90, 120] loss: 0.244 [90, 180] loss: 0.244 [90, 240] loss: 0.245 [90, 300] loss: 0.241 [90, 360] loss: 0.243 Epoch: 90 -> Loss: 0.201568320394 Epoch: 90 -> Test Accuracy: 82.87 [91, 60] loss: 0.244 [91, 120] loss: 0.226 [91, 180] loss: 0.240 [91, 240] loss: 0.252 [91, 300] loss: 0.245 [91, 360] loss: 0.250 Epoch: 91 -> Loss: 0.181910797954 Epoch: 91 -> Test Accuracy: 83.01 [92, 60] loss: 0.249 [92, 120] loss: 0.247 [92, 180] loss: 0.248 [92, 240] loss: 0.235 [92, 300] loss: 0.252 [92, 360] loss: 0.231 Epoch: 92 -> Loss: 0.263022780418 Epoch: 92 -> Test Accuracy: 83.03 [93, 60] loss: 0.243 [93, 120] loss: 0.249 [93, 180] loss: 0.241 [93, 240] loss: 0.245 [93, 300] loss: 0.246 [93, 360] loss: 0.253 Epoch: 93 -> Loss: 0.235726192594 Epoch: 93 -> Test Accuracy: 83.1 [94, 60] loss: 0.240 [94, 120] loss: 0.246 [94, 180] loss: 0.244 [94, 240] loss: 0.239 [94, 300] loss: 0.252 [94, 360] loss: 0.234 Epoch: 94 -> Loss: 0.18844217062 Epoch: 94 -> Test Accuracy: 82.92 [95, 60] loss: 0.234 [95, 120] loss: 0.240 [95, 180] loss: 0.245 [95, 240] loss: 0.244 [95, 300] loss: 0.236 [95, 360] loss: 0.247 Epoch: 95 -> Loss: 0.238631889224 Epoch: 95 -> Test Accuracy: 83.03 [96, 60] loss: 0.234 [96, 120] loss: 0.235 [96, 180] loss: 0.254 [96, 240] loss: 0.242 [96, 300] loss: 0.233 [96, 360] loss: 0.242 Epoch: 96 -> Loss: 0.326489895582 Epoch: 96 -> Test Accuracy: 83.05 [97, 60] loss: 0.232 [97, 120] loss: 0.233 [97, 180] loss: 0.238 [97, 240] loss: 0.247 [97, 300] loss: 0.233 [97, 360] loss: 0.233 Epoch: 97 -> Loss: 0.166981846094 Epoch: 97 -> Test Accuracy: 83.03 [98, 60] loss: 0.254 [98, 120] loss: 0.245 [98, 180] loss: 0.228 [98, 240] loss: 0.236 [98, 300] loss: 0.223 [98, 360] loss: 0.234 Epoch: 98 -> Loss: 0.200633198023 Epoch: 98 -> Test Accuracy: 83.22 [99, 60] loss: 0.239 [99, 120] loss: 0.242 [99, 180] loss: 0.234 [99, 240] loss: 0.234 [99, 300] loss: 0.232 [99, 360] loss: 0.237 Epoch: 99 -> Loss: 0.202944040298 Epoch: 99 -> Test Accuracy: 83.3 [100, 60] loss: 0.239 [100, 120] loss: 0.232 [100, 180] loss: 0.238 [100, 240] loss: 0.232 [100, 300] loss: 0.231 [100, 360] loss: 0.242 Epoch: 100 -> Loss: 0.19259929657 Epoch: 100 -> Test Accuracy: 83.12 Finished Training [1, 60] loss: 1.694 [1, 120] loss: 0.854 [1, 180] loss: 0.739 [1, 240] loss: 0.710 [1, 300] loss: 0.680 [1, 360] loss: 0.640 Epoch: 1 -> Loss: 0.636373102665 Epoch: 1 -> Test Accuracy: 79.01 [2, 60] loss: 0.574 [2, 120] loss: 0.599 [2, 180] loss: 0.579 [2, 240] loss: 0.561 [2, 300] loss: 0.572 [2, 360] loss: 0.563 Epoch: 2 -> Loss: 0.623220920563 Epoch: 2 -> Test Accuracy: 80.69 [3, 60] loss: 0.516 [3, 120] loss: 0.521 [3, 180] loss: 0.526 [3, 240] loss: 0.504 [3, 300] loss: 0.506 [3, 360] loss: 0.519 Epoch: 3 -> Loss: 0.387173980474 Epoch: 3 -> Test Accuracy: 81.6 [4, 60] loss: 0.482 [4, 120] loss: 0.480 [4, 180] loss: 0.460 [4, 240] loss: 0.482 [4, 300] loss: 0.479 [4, 360] loss: 0.472 Epoch: 4 -> Loss: 0.515356242657 Epoch: 4 -> Test Accuracy: 83.1 [5, 60] loss: 0.448 [5, 120] loss: 0.443 [5, 180] loss: 0.442 [5, 240] loss: 0.449 [5, 300] loss: 0.459 [5, 360] loss: 0.454 Epoch: 5 -> Loss: 0.620316267014 Epoch: 5 -> Test Accuracy: 82.64 [6, 60] loss: 0.433 [6, 120] loss: 0.421 [6, 180] loss: 0.452 [6, 240] loss: 0.436 [6, 300] loss: 0.440 [6, 360] loss: 0.444 Epoch: 6 -> Loss: 0.475322335958 Epoch: 6 -> Test Accuracy: 82.88 [7, 60] loss: 0.411 [7, 120] loss: 0.427 [7, 180] loss: 0.422 [7, 240] loss: 0.420 [7, 300] loss: 0.433 [7, 360] loss: 0.438 Epoch: 7 -> Loss: 0.601088821888 Epoch: 7 -> Test Accuracy: 83.59 [8, 60] loss: 0.430 [8, 120] loss: 0.398 [8, 180] loss: 0.408 [8, 240] loss: 0.416 [8, 300] loss: 0.419 [8, 360] loss: 0.431 Epoch: 8 -> Loss: 0.451033890247 Epoch: 8 -> Test Accuracy: 83.35 [9, 60] loss: 0.388 [9, 120] loss: 0.391 [9, 180] loss: 0.421 [9, 240] loss: 0.417 [9, 300] loss: 0.428 [9, 360] loss: 0.423 Epoch: 9 -> Loss: 0.294439375401 Epoch: 9 -> Test Accuracy: 83.37 [10, 60] loss: 0.384 [10, 120] loss: 0.392 [10, 180] loss: 0.407 [10, 240] loss: 0.400 [10, 300] loss: 0.408 [10, 360] loss: 0.416 Epoch: 10 -> Loss: 0.517010211945 Epoch: 10 -> Test Accuracy: 83.87 [11, 60] loss: 0.380 [11, 120] loss: 0.376 [11, 180] loss: 0.406 [11, 240] loss: 0.406 [11, 300] loss: 0.401 [11, 360] loss: 0.411 Epoch: 11 -> Loss: 0.370214641094 Epoch: 11 -> Test Accuracy: 84.18 [12, 60] loss: 0.388 [12, 120] loss: 0.393 [12, 180] loss: 0.393 [12, 240] loss: 0.385 [12, 300] loss: 0.406 [12, 360] loss: 0.388 Epoch: 12 -> Loss: 0.292529672384 Epoch: 12 -> Test Accuracy: 84.02 [13, 60] loss: 0.374 [13, 120] loss: 0.389 [13, 180] loss: 0.379 [13, 240] loss: 0.412 [13, 300] loss: 0.384 [13, 360] loss: 0.383 Epoch: 13 -> Loss: 0.447354316711 Epoch: 13 -> Test Accuracy: 83.77 [14, 60] loss: 0.348 [14, 120] loss: 0.391 [14, 180] loss: 0.389 [14, 240] loss: 0.394 [14, 300] loss: 0.398 [14, 360] loss: 0.398 Epoch: 14 -> Loss: 0.337503701448 Epoch: 14 -> Test Accuracy: 83.95 [15, 60] loss: 0.355 [15, 120] loss: 0.364 [15, 180] loss: 0.384 [15, 240] loss: 0.395 [15, 300] loss: 0.379 [15, 360] loss: 0.397 Epoch: 15 -> Loss: 0.357359498739 Epoch: 15 -> Test Accuracy: 84.22 [16, 60] loss: 0.360 [16, 120] loss: 0.380 [16, 180] loss: 0.368 [16, 240] loss: 0.387 [16, 300] loss: 0.373 [16, 360] loss: 0.403 Epoch: 16 -> Loss: 0.50570756197 Epoch: 16 -> Test Accuracy: 84.19 [17, 60] loss: 0.354 [17, 120] loss: 0.354 [17, 180] loss: 0.385 [17, 240] loss: 0.372 [17, 300] loss: 0.394 [17, 360] loss: 0.402 Epoch: 17 -> Loss: 0.459813982248 Epoch: 17 -> Test Accuracy: 83.83 [18, 60] loss: 0.346 [18, 120] loss: 0.362 [18, 180] loss: 0.375 [18, 240] loss: 0.378 [18, 300] loss: 0.392 [18, 360] loss: 0.377 Epoch: 18 -> Loss: 0.312015414238 Epoch: 18 -> Test Accuracy: 83.88 [19, 60] loss: 0.359 [19, 120] loss: 0.364 [19, 180] loss: 0.369 [19, 240] loss: 0.379 [19, 300] loss: 0.374 [19, 360] loss: 0.387 Epoch: 19 -> Loss: 0.488042891026 Epoch: 19 -> Test Accuracy: 83.76 [20, 60] loss: 0.348 [20, 120] loss: 0.375 [20, 180] loss: 0.365 [20, 240] loss: 0.376 [20, 300] loss: 0.371 [20, 360] loss: 0.403 Epoch: 20 -> Loss: 0.45441275835 Epoch: 20 -> Test Accuracy: 84.08 [21, 60] loss: 0.325 [21, 120] loss: 0.298 [21, 180] loss: 0.300 [21, 240] loss: 0.286 [21, 300] loss: 0.295 [21, 360] loss: 0.271 Epoch: 21 -> Loss: 0.223467111588 Epoch: 21 -> Test Accuracy: 86.25 [22, 60] loss: 0.252 [22, 120] loss: 0.256 [22, 180] loss: 0.279 [22, 240] loss: 0.269 [22, 300] loss: 0.258 [22, 360] loss: 0.248 Epoch: 22 -> Loss: 0.456174194813 Epoch: 22 -> Test Accuracy: 86.18 [23, 60] loss: 0.250 [23, 120] loss: 0.243 [23, 180] loss: 0.254 [23, 240] loss: 0.240 [23, 300] loss: 0.257 [23, 360] loss: 0.256 Epoch: 23 -> Loss: 0.266110479832 Epoch: 23 -> Test Accuracy: 86.31 [24, 60] loss: 0.231 [24, 120] loss: 0.232 [24, 180] loss: 0.230 [24, 240] loss: 0.256 [24, 300] loss: 0.232 [24, 360] loss: 0.242 Epoch: 24 -> Loss: 0.224151775241 Epoch: 24 -> Test Accuracy: 86.27 [25, 60] loss: 0.227 [25, 120] loss: 0.220 [25, 180] loss: 0.232 [25, 240] loss: 0.215 [25, 300] loss: 0.238 [25, 360] loss: 0.231 Epoch: 25 -> Loss: 0.272114276886 Epoch: 25 -> Test Accuracy: 86.17 [26, 60] loss: 0.204 [26, 120] loss: 0.209 [26, 180] loss: 0.221 [26, 240] loss: 0.216 [26, 300] loss: 0.228 [26, 360] loss: 0.235 Epoch: 26 -> Loss: 0.226827740669 Epoch: 26 -> Test Accuracy: 86.28 [27, 60] loss: 0.222 [27, 120] loss: 0.220 [27, 180] loss: 0.218 [27, 240] loss: 0.210 [27, 300] loss: 0.214 [27, 360] loss: 0.232 Epoch: 27 -> Loss: 0.18773420155 Epoch: 27 -> Test Accuracy: 86.02 [28, 60] loss: 0.207 [28, 120] loss: 0.215 [28, 180] loss: 0.223 [28, 240] loss: 0.207 [28, 300] loss: 0.220 [28, 360] loss: 0.220 Epoch: 28 -> Loss: 0.34678658843 Epoch: 28 -> Test Accuracy: 85.59 [29, 60] loss: 0.191 [29, 120] loss: 0.216 [29, 180] loss: 0.225 [29, 240] loss: 0.210 [29, 300] loss: 0.204 [29, 360] loss: 0.213 Epoch: 29 -> Loss: 0.256274312735 Epoch: 29 -> Test Accuracy: 85.9 [30, 60] loss: 0.193 [30, 120] loss: 0.201 [30, 180] loss: 0.209 [30, 240] loss: 0.210 [30, 300] loss: 0.202 [30, 360] loss: 0.208 Epoch: 30 -> Loss: 0.269257485867 Epoch: 30 -> Test Accuracy: 85.51 [31, 60] loss: 0.192 [31, 120] loss: 0.196 [31, 180] loss: 0.199 [31, 240] loss: 0.203 [31, 300] loss: 0.206 [31, 360] loss: 0.218 Epoch: 31 -> Loss: 0.22524318099 Epoch: 31 -> Test Accuracy: 85.74 [32, 60] loss: 0.199 [32, 120] loss: 0.196 [32, 180] loss: 0.195 [32, 240] loss: 0.197 [32, 300] loss: 0.205 [32, 360] loss: 0.204 Epoch: 32 -> Loss: 0.305242300034 Epoch: 32 -> Test Accuracy: 85.9 [33, 60] loss: 0.192 [33, 120] loss: 0.209 [33, 180] loss: 0.216 [33, 240] loss: 0.191 [33, 300] loss: 0.209 [33, 360] loss: 0.206 Epoch: 33 -> Loss: 0.265587151051 Epoch: 33 -> Test Accuracy: 85.81 [34, 60] loss: 0.203 [34, 120] loss: 0.199 [34, 180] loss: 0.213 [34, 240] loss: 0.204 [34, 300] loss: 0.208 [34, 360] loss: 0.202 Epoch: 34 -> Loss: 0.298242270947 Epoch: 34 -> Test Accuracy: 85.5 [35, 60] loss: 0.191 [35, 120] loss: 0.201 [35, 180] loss: 0.204 [35, 240] loss: 0.199 [35, 300] loss: 0.209 [35, 360] loss: 0.217 Epoch: 35 -> Loss: 0.330903589725 Epoch: 35 -> Test Accuracy: 85.81 [36, 60] loss: 0.193 [36, 120] loss: 0.200 [36, 180] loss: 0.198 [36, 240] loss: 0.197 [36, 300] loss: 0.202 [36, 360] loss: 0.204 Epoch: 36 -> Loss: 0.358889877796 Epoch: 36 -> Test Accuracy: 85.75 [37, 60] loss: 0.186 [37, 120] loss: 0.189 [37, 180] loss: 0.195 [37, 240] loss: 0.210 [37, 300] loss: 0.193 [37, 360] loss: 0.205 Epoch: 37 -> Loss: 0.110308326781 Epoch: 37 -> Test Accuracy: 85.42 [38, 60] loss: 0.182 [38, 120] loss: 0.185 [38, 180] loss: 0.187 [38, 240] loss: 0.205 [38, 300] loss: 0.197 [38, 360] loss: 0.204 Epoch: 38 -> Loss: 0.329784154892 Epoch: 38 -> Test Accuracy: 85.51 [39, 60] loss: 0.196 [39, 120] loss: 0.190 [39, 180] loss: 0.189 [39, 240] loss: 0.185 [39, 300] loss: 0.197 [39, 360] loss: 0.204 Epoch: 39 -> Loss: 0.115293264389 Epoch: 39 -> Test Accuracy: 85.29 [40, 60] loss: 0.190 [40, 120] loss: 0.180 [40, 180] loss: 0.200 [40, 240] loss: 0.203 [40, 300] loss: 0.201 [40, 360] loss: 0.206 Epoch: 40 -> Loss: 0.194905668497 Epoch: 40 -> Test Accuracy: 85.65 [41, 60] loss: 0.173 [41, 120] loss: 0.170 [41, 180] loss: 0.160 [41, 240] loss: 0.168 [41, 300] loss: 0.168 [41, 360] loss: 0.159 Epoch: 41 -> Loss: 0.138016298413 Epoch: 41 -> Test Accuracy: 86.18 [42, 60] loss: 0.155 [42, 120] loss: 0.152 [42, 180] loss: 0.153 [42, 240] loss: 0.140 [42, 300] loss: 0.148 [42, 360] loss: 0.151 Epoch: 42 -> Loss: 0.0941276699305 Epoch: 42 -> Test Accuracy: 86.37 [43, 60] loss: 0.128 [43, 120] loss: 0.136 [43, 180] loss: 0.145 [43, 240] loss: 0.135 [43, 300] loss: 0.141 [43, 360] loss: 0.137 Epoch: 43 -> Loss: 0.142272397876 Epoch: 43 -> Test Accuracy: 86.57 [44, 60] loss: 0.133 [44, 120] loss: 0.128 [44, 180] loss: 0.118 [44, 240] loss: 0.131 [44, 300] loss: 0.136 [44, 360] loss: 0.141 Epoch: 44 -> Loss: 0.19723905623 Epoch: 44 -> Test Accuracy: 86.43 [45, 60] loss: 0.125 [45, 120] loss: 0.128 [45, 180] loss: 0.131 [45, 240] loss: 0.125 [45, 300] loss: 0.126 [45, 360] loss: 0.124 Epoch: 45 -> Loss: 0.06195602566 Epoch: 45 -> Test Accuracy: 86.59 [46, 60] loss: 0.120 [46, 120] loss: 0.117 [46, 180] loss: 0.127 [46, 240] loss: 0.120 [46, 300] loss: 0.124 [46, 360] loss: 0.125 Epoch: 46 -> Loss: 0.147839203477 Epoch: 46 -> Test Accuracy: 86.53 [47, 60] loss: 0.121 [47, 120] loss: 0.107 [47, 180] loss: 0.106 [47, 240] loss: 0.121 [47, 300] loss: 0.125 [47, 360] loss: 0.123 Epoch: 47 -> Loss: 0.202754691243 Epoch: 47 -> Test Accuracy: 86.58 [48, 60] loss: 0.123 [48, 120] loss: 0.104 [48, 180] loss: 0.114 [48, 240] loss: 0.117 [48, 300] loss: 0.127 [48, 360] loss: 0.116 Epoch: 48 -> Loss: 0.117970824242 Epoch: 48 -> Test Accuracy: 86.55 [49, 60] loss: 0.120 [49, 120] loss: 0.116 [49, 180] loss: 0.109 [49, 240] loss: 0.122 [49, 300] loss: 0.118 [49, 360] loss: 0.119 Epoch: 49 -> Loss: 0.12256872654 Epoch: 49 -> Test Accuracy: 86.55 [50, 60] loss: 0.120 [50, 120] loss: 0.116 [50, 180] loss: 0.115 [50, 240] loss: 0.121 [50, 300] loss: 0.108 [50, 360] loss: 0.114 Epoch: 50 -> Loss: 0.0400307402015 Epoch: 50 -> Test Accuracy: 86.63 [51, 60] loss: 0.117 [51, 120] loss: 0.102 [51, 180] loss: 0.111 [51, 240] loss: 0.114 [51, 300] loss: 0.111 [51, 360] loss: 0.107 Epoch: 51 -> Loss: 0.0920950621367 Epoch: 51 -> Test Accuracy: 86.6 [52, 60] loss: 0.113 [52, 120] loss: 0.106 [52, 180] loss: 0.106 [52, 240] loss: 0.112 [52, 300] loss: 0.115 [52, 360] loss: 0.114 Epoch: 52 -> Loss: 0.114840552211 Epoch: 52 -> Test Accuracy: 86.66 [53, 60] loss: 0.111 [53, 120] loss: 0.101 [53, 180] loss: 0.108 [53, 240] loss: 0.114 [53, 300] loss: 0.115 [53, 360] loss: 0.102 Epoch: 53 -> Loss: 0.0632527545094 Epoch: 53 -> Test Accuracy: 86.68 [54, 60] loss: 0.105 [54, 120] loss: 0.112 [54, 180] loss: 0.111 [54, 240] loss: 0.113 [54, 300] loss: 0.103 [54, 360] loss: 0.102 Epoch: 54 -> Loss: 0.251395791769 Epoch: 54 -> Test Accuracy: 86.68 [55, 60] loss: 0.107 [55, 120] loss: 0.103 [55, 180] loss: 0.110 [55, 240] loss: 0.105 [55, 300] loss: 0.115 [55, 360] loss: 0.102 Epoch: 55 -> Loss: 0.0681119412184 Epoch: 55 -> Test Accuracy: 86.67 [56, 60] loss: 0.103 [56, 120] loss: 0.105 [56, 180] loss: 0.104 [56, 240] loss: 0.107 [56, 300] loss: 0.111 [56, 360] loss: 0.104 Epoch: 56 -> Loss: 0.149298399687 Epoch: 56 -> Test Accuracy: 86.62 [57, 60] loss: 0.115 [57, 120] loss: 0.105 [57, 180] loss: 0.112 [57, 240] loss: 0.104 [57, 300] loss: 0.107 [57, 360] loss: 0.102 Epoch: 57 -> Loss: 0.127135068178 Epoch: 57 -> Test Accuracy: 86.62 [58, 60] loss: 0.105 [58, 120] loss: 0.103 [58, 180] loss: 0.102 [58, 240] loss: 0.104 [58, 300] loss: 0.103 [58, 360] loss: 0.102 Epoch: 58 -> Loss: 0.0885294824839 Epoch: 58 -> Test Accuracy: 86.71 [59, 60] loss: 0.105 [59, 120] loss: 0.109 [59, 180] loss: 0.103 [59, 240] loss: 0.108 [59, 300] loss: 0.103 [59, 360] loss: 0.099 Epoch: 59 -> Loss: 0.122068703175 Epoch: 59 -> Test Accuracy: 86.67 [60, 60] loss: 0.104 [60, 120] loss: 0.109 [60, 180] loss: 0.100 [60, 240] loss: 0.099 [60, 300] loss: 0.101 [60, 360] loss: 0.105 Epoch: 60 -> Loss: 0.0522882454097 Epoch: 60 -> Test Accuracy: 86.72 [61, 60] loss: 0.104 [61, 120] loss: 0.105 [61, 180] loss: 0.104 [61, 240] loss: 0.101 [61, 300] loss: 0.100 [61, 360] loss: 0.104 Epoch: 61 -> Loss: 0.0897772461176 Epoch: 61 -> Test Accuracy: 86.79 [62, 60] loss: 0.106 [62, 120] loss: 0.097 [62, 180] loss: 0.099 [62, 240] loss: 0.095 [62, 300] loss: 0.101 [62, 360] loss: 0.098 Epoch: 62 -> Loss: 0.102323554456 Epoch: 62 -> Test Accuracy: 86.58 [63, 60] loss: 0.101 [63, 120] loss: 0.097 [63, 180] loss: 0.094 [63, 240] loss: 0.107 [63, 300] loss: 0.099 [63, 360] loss: 0.097 Epoch: 63 -> Loss: 0.12129150331 Epoch: 63 -> Test Accuracy: 86.56 [64, 60] loss: 0.093 [64, 120] loss: 0.097 [64, 180] loss: 0.098 [64, 240] loss: 0.098 [64, 300] loss: 0.098 [64, 360] loss: 0.097 Epoch: 64 -> Loss: 0.160067588091 Epoch: 64 -> Test Accuracy: 86.48 [65, 60] loss: 0.095 [65, 120] loss: 0.099 [65, 180] loss: 0.100 [65, 240] loss: 0.096 [65, 300] loss: 0.091 [65, 360] loss: 0.095 Epoch: 65 -> Loss: 0.128874734044 Epoch: 65 -> Test Accuracy: 86.46 [66, 60] loss: 0.093 [66, 120] loss: 0.097 [66, 180] loss: 0.098 [66, 240] loss: 0.093 [66, 300] loss: 0.097 [66, 360] loss: 0.102 Epoch: 66 -> Loss: 0.112519145012 Epoch: 66 -> Test Accuracy: 86.64 [67, 60] loss: 0.103 [67, 120] loss: 0.097 [67, 180] loss: 0.101 [67, 240] loss: 0.101 [67, 300] loss: 0.096 [67, 360] loss: 0.096 Epoch: 67 -> Loss: 0.145581796765 Epoch: 67 -> Test Accuracy: 86.67 [68, 60] loss: 0.096 [68, 120] loss: 0.093 [68, 180] loss: 0.096 [68, 240] loss: 0.097 [68, 300] loss: 0.090 [68, 360] loss: 0.102 Epoch: 68 -> Loss: 0.0466391667724 Epoch: 68 -> Test Accuracy: 86.58 [69, 60] loss: 0.092 [69, 120] loss: 0.094 [69, 180] loss: 0.100 [69, 240] loss: 0.091 [69, 300] loss: 0.101 [69, 360] loss: 0.096 Epoch: 69 -> Loss: 0.125960662961 Epoch: 69 -> Test Accuracy: 86.5 [70, 60] loss: 0.090 [70, 120] loss: 0.097 [70, 180] loss: 0.096 [70, 240] loss: 0.102 [70, 300] loss: 0.098 [70, 360] loss: 0.098 Epoch: 70 -> Loss: 0.143998235464 Epoch: 70 -> Test Accuracy: 86.55 [71, 60] loss: 0.092 [71, 120] loss: 0.103 [71, 180] loss: 0.095 [71, 240] loss: 0.097 [71, 300] loss: 0.088 [71, 360] loss: 0.092 Epoch: 71 -> Loss: 0.0564641170204 Epoch: 71 -> Test Accuracy: 86.44 [72, 60] loss: 0.095 [72, 120] loss: 0.084 [72, 180] loss: 0.103 [72, 240] loss: 0.090 [72, 300] loss: 0.098 [72, 360] loss: 0.096 Epoch: 72 -> Loss: 0.122200034559 Epoch: 72 -> Test Accuracy: 86.66 [73, 60] loss: 0.089 [73, 120] loss: 0.091 [73, 180] loss: 0.089 [73, 240] loss: 0.095 [73, 300] loss: 0.094 [73, 360] loss: 0.089 Epoch: 73 -> Loss: 0.175251856446 Epoch: 73 -> Test Accuracy: 86.49 [74, 60] loss: 0.089 [74, 120] loss: 0.088 [74, 180] loss: 0.088 [74, 240] loss: 0.093 [74, 300] loss: 0.096 [74, 360] loss: 0.092 Epoch: 74 -> Loss: 0.109243236482 Epoch: 74 -> Test Accuracy: 86.51 [75, 60] loss: 0.091 [75, 120] loss: 0.089 [75, 180] loss: 0.086 [75, 240] loss: 0.093 [75, 300] loss: 0.094 [75, 360] loss: 0.084 Epoch: 75 -> Loss: 0.157480970025 Epoch: 75 -> Test Accuracy: 86.42 [76, 60] loss: 0.088 [76, 120] loss: 0.088 [76, 180] loss: 0.092 [76, 240] loss: 0.087 [76, 300] loss: 0.094 [76, 360] loss: 0.096 Epoch: 76 -> Loss: 0.143909007311 Epoch: 76 -> Test Accuracy: 86.48 [77, 60] loss: 0.087 [77, 120] loss: 0.093 [77, 180] loss: 0.088 [77, 240] loss: 0.086 [77, 300] loss: 0.087 [77, 360] loss: 0.088 Epoch: 77 -> Loss: 0.0574901998043 Epoch: 77 -> Test Accuracy: 86.53 [78, 60] loss: 0.092 [78, 120] loss: 0.099 [78, 180] loss: 0.089 [78, 240] loss: 0.090 [78, 300] loss: 0.090 [78, 360] loss: 0.087 Epoch: 78 -> Loss: 0.195210769773 Epoch: 78 -> Test Accuracy: 86.44 [79, 60] loss: 0.089 [79, 120] loss: 0.085 [79, 180] loss: 0.090 [79, 240] loss: 0.087 [79, 300] loss: 0.089 [79, 360] loss: 0.084 Epoch: 79 -> Loss: 0.0632743090391 Epoch: 79 -> Test Accuracy: 86.65 [80, 60] loss: 0.086 [80, 120] loss: 0.091 [80, 180] loss: 0.085 [80, 240] loss: 0.088 [80, 300] loss: 0.089 [80, 360] loss: 0.084 Epoch: 80 -> Loss: 0.0598878189921 Epoch: 80 -> Test Accuracy: 86.67 [81, 60] loss: 0.094 [81, 120] loss: 0.080 [81, 180] loss: 0.079 [81, 240] loss: 0.087 [81, 300] loss: 0.084 [81, 360] loss: 0.088 Epoch: 81 -> Loss: 0.0422129034996 Epoch: 81 -> Test Accuracy: 86.62 [82, 60] loss: 0.081 [82, 120] loss: 0.084 [82, 180] loss: 0.088 [82, 240] loss: 0.085 [82, 300] loss: 0.087 [82, 360] loss: 0.090 Epoch: 82 -> Loss: 0.220996469259 Epoch: 82 -> Test Accuracy: 86.54 [83, 60] loss: 0.091 [83, 120] loss: 0.091 [83, 180] loss: 0.083 [83, 240] loss: 0.087 [83, 300] loss: 0.086 [83, 360] loss: 0.087 Epoch: 83 -> Loss: 0.0822087526321 Epoch: 83 -> Test Accuracy: 86.47 [84, 60] loss: 0.085 [84, 120] loss: 0.084 [84, 180] loss: 0.090 [84, 240] loss: 0.088 [84, 300] loss: 0.087 [84, 360] loss: 0.096 Epoch: 84 -> Loss: 0.0934516340494 Epoch: 84 -> Test Accuracy: 86.48 [85, 60] loss: 0.088 [85, 120] loss: 0.090 [85, 180] loss: 0.092 [85, 240] loss: 0.082 [85, 300] loss: 0.088 [85, 360] loss: 0.082 Epoch: 85 -> Loss: 0.0341459698975 Epoch: 85 -> Test Accuracy: 86.39 [86, 60] loss: 0.086 [86, 120] loss: 0.086 [86, 180] loss: 0.086 [86, 240] loss: 0.090 [86, 300] loss: 0.079 [86, 360] loss: 0.086 Epoch: 86 -> Loss: 0.0619657225907 Epoch: 86 -> Test Accuracy: 86.41 [87, 60] loss: 0.080 [87, 120] loss: 0.081 [87, 180] loss: 0.085 [87, 240] loss: 0.089 [87, 300] loss: 0.090 [87, 360] loss: 0.086 Epoch: 87 -> Loss: 0.102124616504 Epoch: 87 -> Test Accuracy: 86.45 [88, 60] loss: 0.080 [88, 120] loss: 0.085 [88, 180] loss: 0.074 [88, 240] loss: 0.082 [88, 300] loss: 0.085 [88, 360] loss: 0.075 Epoch: 88 -> Loss: 0.0687177181244 Epoch: 88 -> Test Accuracy: 86.45 [89, 60] loss: 0.084 [89, 120] loss: 0.088 [89, 180] loss: 0.080 [89, 240] loss: 0.085 [89, 300] loss: 0.079 [89, 360] loss: 0.080 Epoch: 89 -> Loss: 0.0518793687224 Epoch: 89 -> Test Accuracy: 86.44 [90, 60] loss: 0.080 [90, 120] loss: 0.083 [90, 180] loss: 0.077 [90, 240] loss: 0.075 [90, 300] loss: 0.080 [90, 360] loss: 0.086 Epoch: 90 -> Loss: 0.0901532620192 Epoch: 90 -> Test Accuracy: 86.43 [91, 60] loss: 0.076 [91, 120] loss: 0.082 [91, 180] loss: 0.085 [91, 240] loss: 0.083 [91, 300] loss: 0.081 [91, 360] loss: 0.079 Epoch: 91 -> Loss: 0.0522777028382 Epoch: 91 -> Test Accuracy: 86.52 [92, 60] loss: 0.074 [92, 120] loss: 0.081 [92, 180] loss: 0.074 [92, 240] loss: 0.084 [92, 300] loss: 0.079 [92, 360] loss: 0.081 Epoch: 92 -> Loss: 0.0731507614255 Epoch: 92 -> Test Accuracy: 86.48 [93, 60] loss: 0.083 [93, 120] loss: 0.078 [93, 180] loss: 0.073 [93, 240] loss: 0.080 [93, 300] loss: 0.076 [93, 360] loss: 0.082 Epoch: 93 -> Loss: 0.129418283701 Epoch: 93 -> Test Accuracy: 86.52 [94, 60] loss: 0.084 [94, 120] loss: 0.080 [94, 180] loss: 0.078 [94, 240] loss: 0.082 [94, 300] loss: 0.084 [94, 360] loss: 0.076 Epoch: 94 -> Loss: 0.0669811069965 Epoch: 94 -> Test Accuracy: 86.49 [95, 60] loss: 0.078 [95, 120] loss: 0.079 [95, 180] loss: 0.087 [95, 240] loss: 0.078 [95, 300] loss: 0.072 [95, 360] loss: 0.079 Epoch: 95 -> Loss: 0.0709091946483 Epoch: 95 -> Test Accuracy: 86.5 [96, 60] loss: 0.076 [96, 120] loss: 0.073 [96, 180] loss: 0.079 [96, 240] loss: 0.082 [96, 300] loss: 0.080 [96, 360] loss: 0.085 Epoch: 96 -> Loss: 0.0867193639278 Epoch: 96 -> Test Accuracy: 86.54 [97, 60] loss: 0.083 [97, 120] loss: 0.084 [97, 180] loss: 0.075 [97, 240] loss: 0.077 [97, 300] loss: 0.076 [97, 360] loss: 0.073 Epoch: 97 -> Loss: 0.0352884605527 Epoch: 97 -> Test Accuracy: 86.65 [98, 60] loss: 0.081 [98, 120] loss: 0.081 [98, 180] loss: 0.084 [98, 240] loss: 0.075 [98, 300] loss: 0.074 [98, 360] loss: 0.077 Epoch: 98 -> Loss: 0.0382921583951 Epoch: 98 -> Test Accuracy: 86.66 [99, 60] loss: 0.076 [99, 120] loss: 0.084 [99, 180] loss: 0.078 [99, 240] loss: 0.078 [99, 300] loss: 0.073 [99, 360] loss: 0.081 Epoch: 99 -> Loss: 0.0843428224325 Epoch: 99 -> Test Accuracy: 86.66 [100, 60] loss: 0.074 [100, 120] loss: 0.080 [100, 180] loss: 0.077 [100, 240] loss: 0.077 [100, 300] loss: 0.078 [100, 360] loss: 0.080 Epoch: 100 -> Loss: 0.116184517741 Epoch: 100 -> Test Accuracy: 86.6 Finished Training [1, 60] loss: 1.597 [1, 120] loss: 0.856 [1, 180] loss: 0.815 [1, 240] loss: 0.771 [1, 300] loss: 0.732 [1, 360] loss: 0.709 Epoch: 1 -> Loss: 0.512435734272 Epoch: 1 -> Test Accuracy: 74.67 [2, 60] loss: 0.662 [2, 120] loss: 0.654 [2, 180] loss: 0.652 [2, 240] loss: 0.652 [2, 300] loss: 0.628 [2, 360] loss: 0.628 Epoch: 2 -> Loss: 0.862444281578 Epoch: 2 -> Test Accuracy: 77.04 [3, 60] loss: 0.599 [3, 120] loss: 0.591 [3, 180] loss: 0.597 [3, 240] loss: 0.597 [3, 300] loss: 0.584 [3, 360] loss: 0.591 Epoch: 3 -> Loss: 0.437123596668 Epoch: 3 -> Test Accuracy: 76.87 [4, 60] loss: 0.566 [4, 120] loss: 0.564 [4, 180] loss: 0.562 [4, 240] loss: 0.540 [4, 300] loss: 0.557 [4, 360] loss: 0.562 Epoch: 4 -> Loss: 0.742728292942 Epoch: 4 -> Test Accuracy: 78.66 [5, 60] loss: 0.543 [5, 120] loss: 0.561 [5, 180] loss: 0.548 [5, 240] loss: 0.532 [5, 300] loss: 0.544 [5, 360] loss: 0.529 Epoch: 5 -> Loss: 0.580091655254 Epoch: 5 -> Test Accuracy: 78.76 [6, 60] loss: 0.514 [6, 120] loss: 0.527 [6, 180] loss: 0.532 [6, 240] loss: 0.527 [6, 300] loss: 0.523 [6, 360] loss: 0.520 Epoch: 6 -> Loss: 0.561300098896 Epoch: 6 -> Test Accuracy: 78.89 [7, 60] loss: 0.481 [7, 120] loss: 0.510 [7, 180] loss: 0.518 [7, 240] loss: 0.510 [7, 300] loss: 0.522 [7, 360] loss: 0.530 Epoch: 7 -> Loss: 0.464393049479 Epoch: 7 -> Test Accuracy: 78.64 [8, 60] loss: 0.495 [8, 120] loss: 0.504 [8, 180] loss: 0.513 [8, 240] loss: 0.502 [8, 300] loss: 0.507 [8, 360] loss: 0.522 Epoch: 8 -> Loss: 0.66817688942 Epoch: 8 -> Test Accuracy: 79.35 [9, 60] loss: 0.467 [9, 120] loss: 0.507 [9, 180] loss: 0.508 [9, 240] loss: 0.508 [9, 300] loss: 0.501 [9, 360] loss: 0.512 Epoch: 9 -> Loss: 0.687930226326 Epoch: 9 -> Test Accuracy: 79.51 [10, 60] loss: 0.487 [10, 120] loss: 0.499 [10, 180] loss: 0.511 [10, 240] loss: 0.499 [10, 300] loss: 0.499 [10, 360] loss: 0.503 Epoch: 10 -> Loss: 0.542013049126 Epoch: 10 -> Test Accuracy: 79.69 [11, 60] loss: 0.495 [11, 120] loss: 0.491 [11, 180] loss: 0.496 [11, 240] loss: 0.497 [11, 300] loss: 0.517 [11, 360] loss: 0.481 Epoch: 11 -> Loss: 0.486979961395 Epoch: 11 -> Test Accuracy: 79.8 [12, 60] loss: 0.475 [12, 120] loss: 0.486 [12, 180] loss: 0.468 [12, 240] loss: 0.490 [12, 300] loss: 0.475 [12, 360] loss: 0.502 Epoch: 12 -> Loss: 0.420037835836 Epoch: 12 -> Test Accuracy: 79.45 [13, 60] loss: 0.476 [13, 120] loss: 0.463 [13, 180] loss: 0.470 [13, 240] loss: 0.486 [13, 300] loss: 0.490 [13, 360] loss: 0.503 Epoch: 13 -> Loss: 0.459505945444 Epoch: 13 -> Test Accuracy: 79.74 [14, 60] loss: 0.470 [14, 120] loss: 0.477 [14, 180] loss: 0.481 [14, 240] loss: 0.474 [14, 300] loss: 0.496 [14, 360] loss: 0.496 Epoch: 14 -> Loss: 0.476970821619 Epoch: 14 -> Test Accuracy: 79.59 [15, 60] loss: 0.468 [15, 120] loss: 0.467 [15, 180] loss: 0.460 [15, 240] loss: 0.480 [15, 300] loss: 0.511 [15, 360] loss: 0.492 Epoch: 15 -> Loss: 0.600821852684 Epoch: 15 -> Test Accuracy: 79.98 [16, 60] loss: 0.467 [16, 120] loss: 0.460 [16, 180] loss: 0.486 [16, 240] loss: 0.469 [16, 300] loss: 0.478 [16, 360] loss: 0.494 Epoch: 16 -> Loss: 0.390962034464 Epoch: 16 -> Test Accuracy: 79.33 [17, 60] loss: 0.454 [17, 120] loss: 0.459 [17, 180] loss: 0.466 [17, 240] loss: 0.482 [17, 300] loss: 0.463 [17, 360] loss: 0.487 Epoch: 17 -> Loss: 0.57130086422 Epoch: 17 -> Test Accuracy: 79.87 [18, 60] loss: 0.457 [18, 120] loss: 0.450 [18, 180] loss: 0.474 [18, 240] loss: 0.481 [18, 300] loss: 0.479 [18, 360] loss: 0.475 Epoch: 18 -> Loss: 0.460366010666 Epoch: 18 -> Test Accuracy: 79.78 [19, 60] loss: 0.441 [19, 120] loss: 0.475 [19, 180] loss: 0.476 [19, 240] loss: 0.482 [19, 300] loss: 0.494 [19, 360] loss: 0.489 Epoch: 19 -> Loss: 0.454329878092 Epoch: 19 -> Test Accuracy: 80.47 [20, 60] loss: 0.454 [20, 120] loss: 0.476 [20, 180] loss: 0.468 [20, 240] loss: 0.476 [20, 300] loss: 0.460 [20, 360] loss: 0.477 Epoch: 20 -> Loss: 0.285374134779 Epoch: 20 -> Test Accuracy: 79.99 [21, 60] loss: 0.422 [21, 120] loss: 0.410 [21, 180] loss: 0.395 [21, 240] loss: 0.400 [21, 300] loss: 0.384 [21, 360] loss: 0.395 Epoch: 21 -> Loss: 0.526229023933 Epoch: 21 -> Test Accuracy: 81.88 [22, 60] loss: 0.377 [22, 120] loss: 0.379 [22, 180] loss: 0.364 [22, 240] loss: 0.372 [22, 300] loss: 0.377 [22, 360] loss: 0.359 Epoch: 22 -> Loss: 0.288418501616 Epoch: 22 -> Test Accuracy: 82.26 [23, 60] loss: 0.362 [23, 120] loss: 0.352 [23, 180] loss: 0.360 [23, 240] loss: 0.355 [23, 300] loss: 0.363 [23, 360] loss: 0.371 Epoch: 23 -> Loss: 0.465875476599 Epoch: 23 -> Test Accuracy: 82.3 [24, 60] loss: 0.332 [24, 120] loss: 0.338 [24, 180] loss: 0.340 [24, 240] loss: 0.338 [24, 300] loss: 0.342 [24, 360] loss: 0.358 Epoch: 24 -> Loss: 0.334155619144 Epoch: 24 -> Test Accuracy: 82.27 [25, 60] loss: 0.341 [25, 120] loss: 0.355 [25, 180] loss: 0.323 [25, 240] loss: 0.338 [25, 300] loss: 0.350 [25, 360] loss: 0.344 Epoch: 25 -> Loss: 0.487725168467 Epoch: 25 -> Test Accuracy: 82.25 [26, 60] loss: 0.319 [26, 120] loss: 0.331 [26, 180] loss: 0.340 [26, 240] loss: 0.332 [26, 300] loss: 0.349 [26, 360] loss: 0.333 Epoch: 26 -> Loss: 0.412365913391 Epoch: 26 -> Test Accuracy: 82.36 [27, 60] loss: 0.320 [27, 120] loss: 0.332 [27, 180] loss: 0.323 [27, 240] loss: 0.321 [27, 300] loss: 0.336 [27, 360] loss: 0.344 Epoch: 27 -> Loss: 0.274124324322 Epoch: 27 -> Test Accuracy: 82.31 [28, 60] loss: 0.318 [28, 120] loss: 0.321 [28, 180] loss: 0.330 [28, 240] loss: 0.333 [28, 300] loss: 0.331 [28, 360] loss: 0.318 Epoch: 28 -> Loss: 0.238350436091 Epoch: 28 -> Test Accuracy: 82.14 [29, 60] loss: 0.315 [29, 120] loss: 0.318 [29, 180] loss: 0.318 [29, 240] loss: 0.304 [29, 300] loss: 0.312 [29, 360] loss: 0.336 Epoch: 29 -> Loss: 0.30804926157 Epoch: 29 -> Test Accuracy: 82.04 [30, 60] loss: 0.304 [30, 120] loss: 0.330 [30, 180] loss: 0.318 [30, 240] loss: 0.328 [30, 300] loss: 0.319 [30, 360] loss: 0.335 Epoch: 30 -> Loss: 0.315720617771 Epoch: 30 -> Test Accuracy: 81.9 [31, 60] loss: 0.316 [31, 120] loss: 0.313 [31, 180] loss: 0.317 [31, 240] loss: 0.326 [31, 300] loss: 0.327 [31, 360] loss: 0.323 Epoch: 31 -> Loss: 0.403066813946 Epoch: 31 -> Test Accuracy: 82.03 [32, 60] loss: 0.308 [32, 120] loss: 0.318 [32, 180] loss: 0.322 [32, 240] loss: 0.314 [32, 300] loss: 0.320 [32, 360] loss: 0.333 Epoch: 32 -> Loss: 0.405086994171 Epoch: 32 -> Test Accuracy: 82.0 [33, 60] loss: 0.310 [33, 120] loss: 0.307 [33, 180] loss: 0.319 [33, 240] loss: 0.329 [33, 300] loss: 0.319 [33, 360] loss: 0.320 Epoch: 33 -> Loss: 0.22546748817 Epoch: 33 -> Test Accuracy: 81.91 [34, 60] loss: 0.308 [34, 120] loss: 0.317 [34, 180] loss: 0.304 [34, 240] loss: 0.311 [34, 300] loss: 0.312 [34, 360] loss: 0.328 Epoch: 34 -> Loss: 0.248193457723 Epoch: 34 -> Test Accuracy: 81.89 [35, 60] loss: 0.301 [35, 120] loss: 0.313 [35, 180] loss: 0.315 [35, 240] loss: 0.326 [35, 300] loss: 0.318 [35, 360] loss: 0.305 Epoch: 35 -> Loss: 0.251711845398 Epoch: 35 -> Test Accuracy: 81.42 [36, 60] loss: 0.298 [36, 120] loss: 0.317 [36, 180] loss: 0.310 [36, 240] loss: 0.315 [36, 300] loss: 0.315 [36, 360] loss: 0.319 Epoch: 36 -> Loss: 0.360121488571 Epoch: 36 -> Test Accuracy: 81.77 [37, 60] loss: 0.306 [37, 120] loss: 0.319 [37, 180] loss: 0.307 [37, 240] loss: 0.308 [37, 300] loss: 0.315 [37, 360] loss: 0.307 Epoch: 37 -> Loss: 0.321190357208 Epoch: 37 -> Test Accuracy: 81.97 [38, 60] loss: 0.289 [38, 120] loss: 0.300 [38, 180] loss: 0.311 [38, 240] loss: 0.299 [38, 300] loss: 0.321 [38, 360] loss: 0.317 Epoch: 38 -> Loss: 0.319918006659 Epoch: 38 -> Test Accuracy: 81.36 [39, 60] loss: 0.297 [39, 120] loss: 0.297 [39, 180] loss: 0.303 [39, 240] loss: 0.312 [39, 300] loss: 0.297 [39, 360] loss: 0.341 Epoch: 39 -> Loss: 0.377610981464 Epoch: 39 -> Test Accuracy: 81.99 [40, 60] loss: 0.306 [40, 120] loss: 0.290 [40, 180] loss: 0.314 [40, 240] loss: 0.307 [40, 300] loss: 0.300 [40, 360] loss: 0.318 Epoch: 40 -> Loss: 0.338285326958 Epoch: 40 -> Test Accuracy: 80.83 [41, 60] loss: 0.285 [41, 120] loss: 0.265 [41, 180] loss: 0.272 [41, 240] loss: 0.279 [41, 300] loss: 0.261 [41, 360] loss: 0.263 Epoch: 41 -> Loss: 0.30100017786 Epoch: 41 -> Test Accuracy: 82.36 [42, 60] loss: 0.257 [42, 120] loss: 0.243 [42, 180] loss: 0.264 [42, 240] loss: 0.252 [42, 300] loss: 0.263 [42, 360] loss: 0.270 Epoch: 42 -> Loss: 0.420700728893 Epoch: 42 -> Test Accuracy: 82.38 [43, 60] loss: 0.258 [43, 120] loss: 0.241 [43, 180] loss: 0.249 [43, 240] loss: 0.242 [43, 300] loss: 0.249 [43, 360] loss: 0.252 Epoch: 43 -> Loss: 0.241235524416 Epoch: 43 -> Test Accuracy: 82.18 [44, 60] loss: 0.234 [44, 120] loss: 0.237 [44, 180] loss: 0.228 [44, 240] loss: 0.245 [44, 300] loss: 0.247 [44, 360] loss: 0.241 Epoch: 44 -> Loss: 0.193565994501 Epoch: 44 -> Test Accuracy: 82.62 [45, 60] loss: 0.223 [45, 120] loss: 0.223 [45, 180] loss: 0.228 [45, 240] loss: 0.244 [45, 300] loss: 0.236 [45, 360] loss: 0.237 Epoch: 45 -> Loss: 0.336680471897 Epoch: 45 -> Test Accuracy: 82.62 [46, 60] loss: 0.216 [46, 120] loss: 0.220 [46, 180] loss: 0.224 [46, 240] loss: 0.227 [46, 300] loss: 0.220 [46, 360] loss: 0.231 Epoch: 46 -> Loss: 0.201660081744 Epoch: 46 -> Test Accuracy: 82.63 [47, 60] loss: 0.217 [47, 120] loss: 0.214 [47, 180] loss: 0.219 [47, 240] loss: 0.222 [47, 300] loss: 0.216 [47, 360] loss: 0.237 Epoch: 47 -> Loss: 0.283549129963 Epoch: 47 -> Test Accuracy: 82.51 [48, 60] loss: 0.217 [48, 120] loss: 0.216 [48, 180] loss: 0.226 [48, 240] loss: 0.229 [48, 300] loss: 0.224 [48, 360] loss: 0.225 Epoch: 48 -> Loss: 0.179385825992 Epoch: 48 -> Test Accuracy: 82.48 [49, 60] loss: 0.222 [49, 120] loss: 0.213 [49, 180] loss: 0.208 [49, 240] loss: 0.219 [49, 300] loss: 0.230 [49, 360] loss: 0.206 Epoch: 49 -> Loss: 0.187668353319 Epoch: 49 -> Test Accuracy: 82.6 [50, 60] loss: 0.232 [50, 120] loss: 0.208 [50, 180] loss: 0.230 [50, 240] loss: 0.224 [50, 300] loss: 0.217 [50, 360] loss: 0.223 Epoch: 50 -> Loss: 0.270434826612 Epoch: 50 -> Test Accuracy: 82.56 [51, 60] loss: 0.231 [51, 120] loss: 0.204 [51, 180] loss: 0.205 [51, 240] loss: 0.228 [51, 300] loss: 0.225 [51, 360] loss: 0.211 Epoch: 51 -> Loss: 0.11189135164 Epoch: 51 -> Test Accuracy: 82.51 [52, 60] loss: 0.223 [52, 120] loss: 0.215 [52, 180] loss: 0.203 [52, 240] loss: 0.211 [52, 300] loss: 0.220 [52, 360] loss: 0.220 Epoch: 52 -> Loss: 0.250686466694 Epoch: 52 -> Test Accuracy: 82.51 [53, 60] loss: 0.198 [53, 120] loss: 0.217 [53, 180] loss: 0.198 [53, 240] loss: 0.213 [53, 300] loss: 0.215 [53, 360] loss: 0.222 Epoch: 53 -> Loss: 0.163546591997 Epoch: 53 -> Test Accuracy: 82.46 [54, 60] loss: 0.211 [54, 120] loss: 0.201 [54, 180] loss: 0.218 [54, 240] loss: 0.211 [54, 300] loss: 0.215 [54, 360] loss: 0.216 Epoch: 54 -> Loss: 0.26966124773 Epoch: 54 -> Test Accuracy: 82.57 [55, 60] loss: 0.207 [55, 120] loss: 0.225 [55, 180] loss: 0.210 [55, 240] loss: 0.214 [55, 300] loss: 0.217 [55, 360] loss: 0.218 Epoch: 55 -> Loss: 0.195465743542 Epoch: 55 -> Test Accuracy: 82.62 [56, 60] loss: 0.213 [56, 120] loss: 0.208 [56, 180] loss: 0.203 [56, 240] loss: 0.224 [56, 300] loss: 0.202 [56, 360] loss: 0.219 Epoch: 56 -> Loss: 0.161485761404 Epoch: 56 -> Test Accuracy: 82.74 [57, 60] loss: 0.204 [57, 120] loss: 0.207 [57, 180] loss: 0.216 [57, 240] loss: 0.207 [57, 300] loss: 0.209 [57, 360] loss: 0.206 Epoch: 57 -> Loss: 0.172239303589 Epoch: 57 -> Test Accuracy: 82.7 [58, 60] loss: 0.206 [58, 120] loss: 0.206 [58, 180] loss: 0.204 [58, 240] loss: 0.209 [58, 300] loss: 0.205 [58, 360] loss: 0.212 Epoch: 58 -> Loss: 0.146751895547 Epoch: 58 -> Test Accuracy: 82.69 [59, 60] loss: 0.204 [59, 120] loss: 0.205 [59, 180] loss: 0.197 [59, 240] loss: 0.210 [59, 300] loss: 0.201 [59, 360] loss: 0.203 Epoch: 59 -> Loss: 0.207809656858 Epoch: 59 -> Test Accuracy: 82.52 [60, 60] loss: 0.210 [60, 120] loss: 0.200 [60, 180] loss: 0.202 [60, 240] loss: 0.207 [60, 300] loss: 0.216 [60, 360] loss: 0.215 Epoch: 60 -> Loss: 0.289859235287 Epoch: 60 -> Test Accuracy: 82.5 [61, 60] loss: 0.210 [61, 120] loss: 0.203 [61, 180] loss: 0.197 [61, 240] loss: 0.198 [61, 300] loss: 0.209 [61, 360] loss: 0.196 Epoch: 61 -> Loss: 0.305670619011 Epoch: 61 -> Test Accuracy: 82.65 [62, 60] loss: 0.200 [62, 120] loss: 0.212 [62, 180] loss: 0.213 [62, 240] loss: 0.202 [62, 300] loss: 0.206 [62, 360] loss: 0.217 Epoch: 62 -> Loss: 0.172732159495 Epoch: 62 -> Test Accuracy: 82.54 [63, 60] loss: 0.198 [63, 120] loss: 0.199 [63, 180] loss: 0.198 [63, 240] loss: 0.208 [63, 300] loss: 0.197 [63, 360] loss: 0.214 Epoch: 63 -> Loss: 0.234139487147 Epoch: 63 -> Test Accuracy: 82.6 [64, 60] loss: 0.194 [64, 120] loss: 0.205 [64, 180] loss: 0.196 [64, 240] loss: 0.206 [64, 300] loss: 0.208 [64, 360] loss: 0.197 Epoch: 64 -> Loss: 0.276698976755 Epoch: 64 -> Test Accuracy: 82.62 [65, 60] loss: 0.202 [65, 120] loss: 0.209 [65, 180] loss: 0.192 [65, 240] loss: 0.191 [65, 300] loss: 0.201 [65, 360] loss: 0.210 Epoch: 65 -> Loss: 0.181875705719 Epoch: 65 -> Test Accuracy: 82.52 [66, 60] loss: 0.191 [66, 120] loss: 0.192 [66, 180] loss: 0.197 [66, 240] loss: 0.200 [66, 300] loss: 0.197 [66, 360] loss: 0.190 Epoch: 66 -> Loss: 0.210397556424 Epoch: 66 -> Test Accuracy: 82.66 [67, 60] loss: 0.190 [67, 120] loss: 0.198 [67, 180] loss: 0.195 [67, 240] loss: 0.203 [67, 300] loss: 0.205 [67, 360] loss: 0.202 Epoch: 67 -> Loss: 0.103144481778 Epoch: 67 -> Test Accuracy: 82.47 [68, 60] loss: 0.191 [68, 120] loss: 0.199 [68, 180] loss: 0.200 [68, 240] loss: 0.196 [68, 300] loss: 0.195 [68, 360] loss: 0.192 Epoch: 68 -> Loss: 0.177246764302 Epoch: 68 -> Test Accuracy: 82.53 [69, 60] loss: 0.194 [69, 120] loss: 0.186 [69, 180] loss: 0.186 [69, 240] loss: 0.202 [69, 300] loss: 0.193 [69, 360] loss: 0.199 Epoch: 69 -> Loss: 0.328603744507 Epoch: 69 -> Test Accuracy: 82.66 [70, 60] loss: 0.188 [70, 120] loss: 0.186 [70, 180] loss: 0.191 [70, 240] loss: 0.193 [70, 300] loss: 0.186 [70, 360] loss: 0.193 Epoch: 70 -> Loss: 0.1943808496 Epoch: 70 -> Test Accuracy: 82.77 [71, 60] loss: 0.196 [71, 120] loss: 0.201 [71, 180] loss: 0.192 [71, 240] loss: 0.193 [71, 300] loss: 0.196 [71, 360] loss: 0.202 Epoch: 71 -> Loss: 0.20616979897 Epoch: 71 -> Test Accuracy: 82.75 [72, 60] loss: 0.201 [72, 120] loss: 0.190 [72, 180] loss: 0.191 [72, 240] loss: 0.194 [72, 300] loss: 0.197 [72, 360] loss: 0.206 Epoch: 72 -> Loss: 0.27528283 Epoch: 72 -> Test Accuracy: 82.67 [73, 60] loss: 0.192 [73, 120] loss: 0.189 [73, 180] loss: 0.192 [73, 240] loss: 0.188 [73, 300] loss: 0.192 [73, 360] loss: 0.186 Epoch: 73 -> Loss: 0.182364612818 Epoch: 73 -> Test Accuracy: 82.62 [74, 60] loss: 0.186 [74, 120] loss: 0.191 [74, 180] loss: 0.188 [74, 240] loss: 0.196 [74, 300] loss: 0.195 [74, 360] loss: 0.189 Epoch: 74 -> Loss: 0.21389195323 Epoch: 74 -> Test Accuracy: 82.76 [75, 60] loss: 0.179 [75, 120] loss: 0.196 [75, 180] loss: 0.191 [75, 240] loss: 0.190 [75, 300] loss: 0.190 [75, 360] loss: 0.196 Epoch: 75 -> Loss: 0.149008527398 Epoch: 75 -> Test Accuracy: 82.75 [76, 60] loss: 0.192 [76, 120] loss: 0.183 [76, 180] loss: 0.197 [76, 240] loss: 0.193 [76, 300] loss: 0.190 [76, 360] loss: 0.189 Epoch: 76 -> Loss: 0.234543561935 Epoch: 76 -> Test Accuracy: 82.56 [77, 60] loss: 0.191 [77, 120] loss: 0.194 [77, 180] loss: 0.181 [77, 240] loss: 0.188 [77, 300] loss: 0.192 [77, 360] loss: 0.183 Epoch: 77 -> Loss: 0.160980522633 Epoch: 77 -> Test Accuracy: 82.49 [78, 60] loss: 0.182 [78, 120] loss: 0.198 [78, 180] loss: 0.183 [78, 240] loss: 0.185 [78, 300] loss: 0.180 [78, 360] loss: 0.184 Epoch: 78 -> Loss: 0.154733732343 Epoch: 78 -> Test Accuracy: 82.64 [79, 60] loss: 0.195 [79, 120] loss: 0.177 [79, 180] loss: 0.188 [79, 240] loss: 0.183 [79, 300] loss: 0.184 [79, 360] loss: 0.203 Epoch: 79 -> Loss: 0.251921266317 Epoch: 79 -> Test Accuracy: 82.6 [80, 60] loss: 0.193 [80, 120] loss: 0.183 [80, 180] loss: 0.174 [80, 240] loss: 0.176 [80, 300] loss: 0.185 [80, 360] loss: 0.190 Epoch: 80 -> Loss: 0.18039470911 Epoch: 80 -> Test Accuracy: 82.61 [81, 60] loss: 0.187 [81, 120] loss: 0.182 [81, 180] loss: 0.181 [81, 240] loss: 0.184 [81, 300] loss: 0.190 [81, 360] loss: 0.197 Epoch: 81 -> Loss: 0.156178563833 Epoch: 81 -> Test Accuracy: 82.62 [82, 60] loss: 0.192 [82, 120] loss: 0.181 [82, 180] loss: 0.177 [82, 240] loss: 0.197 [82, 300] loss: 0.178 [82, 360] loss: 0.182 Epoch: 82 -> Loss: 0.164409220219 Epoch: 82 -> Test Accuracy: 82.59 [83, 60] loss: 0.189 [83, 120] loss: 0.177 [83, 180] loss: 0.176 [83, 240] loss: 0.191 [83, 300] loss: 0.189 [83, 360] loss: 0.183 Epoch: 83 -> Loss: 0.374662697315 Epoch: 83 -> Test Accuracy: 82.76 [84, 60] loss: 0.171 [84, 120] loss: 0.184 [84, 180] loss: 0.190 [84, 240] loss: 0.195 [84, 300] loss: 0.186 [84, 360] loss: 0.183 Epoch: 84 -> Loss: 0.261164724827 Epoch: 84 -> Test Accuracy: 82.69 [85, 60] loss: 0.186 [85, 120] loss: 0.182 [85, 180] loss: 0.188 [85, 240] loss: 0.176 [85, 300] loss: 0.193 [85, 360] loss: 0.189 Epoch: 85 -> Loss: 0.308419048786 Epoch: 85 -> Test Accuracy: 82.6 [86, 60] loss: 0.177 [86, 120] loss: 0.178 [86, 180] loss: 0.180 [86, 240] loss: 0.186 [86, 300] loss: 0.182 [86, 360] loss: 0.190 Epoch: 86 -> Loss: 0.183074206114 Epoch: 86 -> Test Accuracy: 82.72 [87, 60] loss: 0.178 [87, 120] loss: 0.186 [87, 180] loss: 0.188 [87, 240] loss: 0.172 [87, 300] loss: 0.174 [87, 360] loss: 0.184 Epoch: 87 -> Loss: 0.148081868887 Epoch: 87 -> Test Accuracy: 82.61 [88, 60] loss: 0.174 [88, 120] loss: 0.190 [88, 180] loss: 0.190 [88, 240] loss: 0.171 [88, 300] loss: 0.179 [88, 360] loss: 0.185 Epoch: 88 -> Loss: 0.237703084946 Epoch: 88 -> Test Accuracy: 82.58 [89, 60] loss: 0.185 [89, 120] loss: 0.177 [89, 180] loss: 0.174 [89, 240] loss: 0.181 [89, 300] loss: 0.175 [89, 360] loss: 0.180 Epoch: 89 -> Loss: 0.174320682883 Epoch: 89 -> Test Accuracy: 82.74 [90, 60] loss: 0.177 [90, 120] loss: 0.180 [90, 180] loss: 0.172 [90, 240] loss: 0.192 [90, 300] loss: 0.170 [90, 360] loss: 0.171 Epoch: 90 -> Loss: 0.0973794907331 Epoch: 90 -> Test Accuracy: 82.68 [91, 60] loss: 0.169 [91, 120] loss: 0.178 [91, 180] loss: 0.175 [91, 240] loss: 0.176 [91, 300] loss: 0.175 [91, 360] loss: 0.181 Epoch: 91 -> Loss: 0.243659883738 Epoch: 91 -> Test Accuracy: 82.55 [92, 60] loss: 0.178 [92, 120] loss: 0.180 [92, 180] loss: 0.173 [92, 240] loss: 0.173 [92, 300] loss: 0.192 [92, 360] loss: 0.170 Epoch: 92 -> Loss: 0.156590640545 Epoch: 92 -> Test Accuracy: 82.67 [93, 60] loss: 0.170 [93, 120] loss: 0.176 [93, 180] loss: 0.177 [93, 240] loss: 0.161 [93, 300] loss: 0.176 [93, 360] loss: 0.179 Epoch: 93 -> Loss: 0.119468547404 Epoch: 93 -> Test Accuracy: 82.52 [94, 60] loss: 0.164 [94, 120] loss: 0.170 [94, 180] loss: 0.175 [94, 240] loss: 0.175 [94, 300] loss: 0.173 [94, 360] loss: 0.176 Epoch: 94 -> Loss: 0.200928539038 Epoch: 94 -> Test Accuracy: 82.57 [95, 60] loss: 0.175 [95, 120] loss: 0.186 [95, 180] loss: 0.168 [95, 240] loss: 0.172 [95, 300] loss: 0.170 [95, 360] loss: 0.177 Epoch: 95 -> Loss: 0.22417716682 Epoch: 95 -> Test Accuracy: 82.67 [96, 60] loss: 0.175 [96, 120] loss: 0.175 [96, 180] loss: 0.173 [96, 240] loss: 0.167 [96, 300] loss: 0.170 [96, 360] loss: 0.171 Epoch: 96 -> Loss: 0.214863657951 Epoch: 96 -> Test Accuracy: 82.53 [97, 60] loss: 0.174 [97, 120] loss: 0.169 [97, 180] loss: 0.184 [97, 240] loss: 0.171 [97, 300] loss: 0.173 [97, 360] loss: 0.176 Epoch: 97 -> Loss: 0.113119915128 Epoch: 97 -> Test Accuracy: 82.4 [98, 60] loss: 0.182 [98, 120] loss: 0.174 [98, 180] loss: 0.170 [98, 240] loss: 0.169 [98, 300] loss: 0.178 [98, 360] loss: 0.182 Epoch: 98 -> Loss: 0.194167688489 Epoch: 98 -> Test Accuracy: 82.58 [99, 60] loss: 0.175 [99, 120] loss: 0.179 [99, 180] loss: 0.171 [99, 240] loss: 0.174 [99, 300] loss: 0.177 [99, 360] loss: 0.163 Epoch: 99 -> Loss: 0.12676396966 Epoch: 99 -> Test Accuracy: 82.54 [100, 60] loss: 0.174 [100, 120] loss: 0.169 [100, 180] loss: 0.174 [100, 240] loss: 0.175 [100, 300] loss: 0.170 [100, 360] loss: 0.163 Epoch: 100 -> Loss: 0.092672303319 Epoch: 100 -> Test Accuracy: 82.54 Finished Training [1, 60] loss: 2.818 [1, 120] loss: 2.012 [1, 180] loss: 1.942 [1, 240] loss: 1.912 [1, 300] loss: 1.868 [1, 360] loss: 1.851 Epoch: 1 -> Loss: 1.65035951138 Epoch: 1 -> Test Accuracy: 30.72 [2, 60] loss: 1.816 [2, 120] loss: 1.791 [2, 180] loss: 1.794 [2, 240] loss: 1.777 [2, 300] loss: 1.781 [2, 360] loss: 1.757 Epoch: 2 -> Loss: 1.74936366081 Epoch: 2 -> Test Accuracy: 33.34 [3, 60] loss: 1.732 [3, 120] loss: 1.734 [3, 180] loss: 1.727 [3, 240] loss: 1.717 [3, 300] loss: 1.724 [3, 360] loss: 1.714 Epoch: 3 -> Loss: 1.56685900688 Epoch: 3 -> Test Accuracy: 35.05 [4, 60] loss: 1.716 [4, 120] loss: 1.682 [4, 180] loss: 1.696 [4, 240] loss: 1.702 [4, 300] loss: 1.701 [4, 360] loss: 1.699 Epoch: 4 -> Loss: 1.88525998592 Epoch: 4 -> Test Accuracy: 34.81 [5, 60] loss: 1.681 [5, 120] loss: 1.680 [5, 180] loss: 1.707 [5, 240] loss: 1.669 [5, 300] loss: 1.692 [5, 360] loss: 1.683 Epoch: 5 -> Loss: 1.73990762234 Epoch: 5 -> Test Accuracy: 35.49 [6, 60] loss: 1.693 [6, 120] loss: 1.676 [6, 180] loss: 1.670 [6, 240] loss: 1.651 [6, 300] loss: 1.648 [6, 360] loss: 1.670 Epoch: 6 -> Loss: 1.51242232323 Epoch: 6 -> Test Accuracy: 35.59 [7, 60] loss: 1.670 [7, 120] loss: 1.664 [7, 180] loss: 1.655 [7, 240] loss: 1.673 [7, 300] loss: 1.660 [7, 360] loss: 1.656 Epoch: 7 -> Loss: 1.63683533669 Epoch: 7 -> Test Accuracy: 35.55 [8, 60] loss: 1.668 [8, 120] loss: 1.656 [8, 180] loss: 1.650 [8, 240] loss: 1.645 [8, 300] loss: 1.662 [8, 360] loss: 1.640 Epoch: 8 -> Loss: 1.55725729465 Epoch: 8 -> Test Accuracy: 35.92 [9, 60] loss: 1.659 [9, 120] loss: 1.668 [9, 180] loss: 1.640 [9, 240] loss: 1.662 [9, 300] loss: 1.640 [9, 360] loss: 1.653 Epoch: 9 -> Loss: 1.57050871849 Epoch: 9 -> Test Accuracy: 36.05 [10, 60] loss: 1.657 [10, 120] loss: 1.640 [10, 180] loss: 1.647 [10, 240] loss: 1.635 [10, 300] loss: 1.662 [10, 360] loss: 1.644 Epoch: 10 -> Loss: 1.65176618099 Epoch: 10 -> Test Accuracy: 37.38 [11, 60] loss: 1.630 [11, 120] loss: 1.632 [11, 180] loss: 1.644 [11, 240] loss: 1.635 [11, 300] loss: 1.667 [11, 360] loss: 1.619 Epoch: 11 -> Loss: 1.63378334045 Epoch: 11 -> Test Accuracy: 37.26 [12, 60] loss: 1.618 [12, 120] loss: 1.642 [12, 180] loss: 1.622 [12, 240] loss: 1.633 [12, 300] loss: 1.631 [12, 360] loss: 1.650 Epoch: 12 -> Loss: 1.75447976589 Epoch: 12 -> Test Accuracy: 37.32 [13, 60] loss: 1.634 [13, 120] loss: 1.631 [13, 180] loss: 1.637 [13, 240] loss: 1.650 [13, 300] loss: 1.639 [13, 360] loss: 1.622 Epoch: 13 -> Loss: 1.68028259277 Epoch: 13 -> Test Accuracy: 37.32 [14, 60] loss: 1.633 [14, 120] loss: 1.626 [14, 180] loss: 1.619 [14, 240] loss: 1.638 [14, 300] loss: 1.637 [14, 360] loss: 1.651 Epoch: 14 -> Loss: 1.66220152378 Epoch: 14 -> Test Accuracy: 36.28 [15, 60] loss: 1.635 [15, 120] loss: 1.627 [15, 180] loss: 1.626 [15, 240] loss: 1.622 [15, 300] loss: 1.629 [15, 360] loss: 1.659 Epoch: 15 -> Loss: 1.72832965851 Epoch: 15 -> Test Accuracy: 36.58 [16, 60] loss: 1.629 [16, 120] loss: 1.630 [16, 180] loss: 1.630 [16, 240] loss: 1.630 [16, 300] loss: 1.605 [16, 360] loss: 1.620 Epoch: 16 -> Loss: 1.54632163048 Epoch: 16 -> Test Accuracy: 37.75 [17, 60] loss: 1.619 [17, 120] loss: 1.625 [17, 180] loss: 1.629 [17, 240] loss: 1.634 [17, 300] loss: 1.643 [17, 360] loss: 1.622 Epoch: 17 -> Loss: 1.61425685883 Epoch: 17 -> Test Accuracy: 37.55 [18, 60] loss: 1.618 [18, 120] loss: 1.614 [18, 180] loss: 1.623 [18, 240] loss: 1.615 [18, 300] loss: 1.617 [18, 360] loss: 1.623 Epoch: 18 -> Loss: 1.82452487946 Epoch: 18 -> Test Accuracy: 37.74 [19, 60] loss: 1.627 [19, 120] loss: 1.623 [19, 180] loss: 1.620 [19, 240] loss: 1.627 [19, 300] loss: 1.625 [19, 360] loss: 1.615 Epoch: 19 -> Loss: 1.74086797237 Epoch: 19 -> Test Accuracy: 36.92 [20, 60] loss: 1.618 [20, 120] loss: 1.610 [20, 180] loss: 1.611 [20, 240] loss: 1.616 [20, 300] loss: 1.619 [20, 360] loss: 1.615 Epoch: 20 -> Loss: 1.73075580597 Epoch: 20 -> Test Accuracy: 36.39 [21, 60] loss: 1.576 [21, 120] loss: 1.561 [21, 180] loss: 1.534 [21, 240] loss: 1.530 [21, 300] loss: 1.518 [21, 360] loss: 1.525 Epoch: 21 -> Loss: 1.47958314419 Epoch: 21 -> Test Accuracy: 39.53 [22, 60] loss: 1.511 [22, 120] loss: 1.497 [22, 180] loss: 1.529 [22, 240] loss: 1.516 [22, 300] loss: 1.509 [22, 360] loss: 1.509 Epoch: 22 -> Loss: 1.46137690544 Epoch: 22 -> Test Accuracy: 40.37 [23, 60] loss: 1.486 [23, 120] loss: 1.496 [23, 180] loss: 1.522 [23, 240] loss: 1.486 [23, 300] loss: 1.494 [23, 360] loss: 1.477 Epoch: 23 -> Loss: 1.57786989212 Epoch: 23 -> Test Accuracy: 40.34 [24, 60] loss: 1.483 [24, 120] loss: 1.489 [24, 180] loss: 1.478 [24, 240] loss: 1.505 [24, 300] loss: 1.477 [24, 360] loss: 1.494 Epoch: 24 -> Loss: 1.42657911777 Epoch: 24 -> Test Accuracy: 40.61 [25, 60] loss: 1.485 [25, 120] loss: 1.474 [25, 180] loss: 1.490 [25, 240] loss: 1.490 [25, 300] loss: 1.484 [25, 360] loss: 1.489 Epoch: 25 -> Loss: 1.53933930397 Epoch: 25 -> Test Accuracy: 40.45 [26, 60] loss: 1.469 [26, 120] loss: 1.485 [26, 180] loss: 1.472 [26, 240] loss: 1.484 [26, 300] loss: 1.477 [26, 360] loss: 1.487 Epoch: 26 -> Loss: 1.41461062431 Epoch: 26 -> Test Accuracy: 41.02 [27, 60] loss: 1.473 [27, 120] loss: 1.487 [27, 180] loss: 1.482 [27, 240] loss: 1.486 [27, 300] loss: 1.471 [27, 360] loss: 1.498 Epoch: 27 -> Loss: 1.315107584 Epoch: 27 -> Test Accuracy: 41.59 [28, 60] loss: 1.483 [28, 120] loss: 1.469 [28, 180] loss: 1.485 [28, 240] loss: 1.473 [28, 300] loss: 1.488 [28, 360] loss: 1.476 Epoch: 28 -> Loss: 1.40484452248 Epoch: 28 -> Test Accuracy: 41.09 [29, 60] loss: 1.481 [29, 120] loss: 1.469 [29, 180] loss: 1.485 [29, 240] loss: 1.476 [29, 300] loss: 1.470 [29, 360] loss: 1.465 Epoch: 29 -> Loss: 1.45765554905 Epoch: 29 -> Test Accuracy: 41.56 [30, 60] loss: 1.470 [30, 120] loss: 1.472 [30, 180] loss: 1.475 [30, 240] loss: 1.482 [30, 300] loss: 1.483 [30, 360] loss: 1.484 Epoch: 30 -> Loss: 1.46095252037 Epoch: 30 -> Test Accuracy: 41.19 [31, 60] loss: 1.471 [31, 120] loss: 1.462 [31, 180] loss: 1.484 [31, 240] loss: 1.467 [31, 300] loss: 1.474 [31, 360] loss: 1.462 Epoch: 31 -> Loss: 1.70025658607 Epoch: 31 -> Test Accuracy: 40.76 [32, 60] loss: 1.483 [32, 120] loss: 1.468 [32, 180] loss: 1.479 [32, 240] loss: 1.471 [32, 300] loss: 1.484 [32, 360] loss: 1.463 Epoch: 32 -> Loss: 1.44530463219 Epoch: 32 -> Test Accuracy: 41.73 [33, 60] loss: 1.458 [33, 120] loss: 1.446 [33, 180] loss: 1.482 [33, 240] loss: 1.481 [33, 300] loss: 1.483 [33, 360] loss: 1.470 Epoch: 33 -> Loss: 1.26609611511 Epoch: 33 -> Test Accuracy: 41.65 [34, 60] loss: 1.469 [34, 120] loss: 1.482 [34, 180] loss: 1.477 [34, 240] loss: 1.482 [34, 300] loss: 1.470 [34, 360] loss: 1.467 Epoch: 34 -> Loss: 1.47667622566 Epoch: 34 -> Test Accuracy: 41.26 [35, 60] loss: 1.469 [35, 120] loss: 1.471 [35, 180] loss: 1.482 [35, 240] loss: 1.487 [35, 300] loss: 1.472 [35, 360] loss: 1.470 Epoch: 35 -> Loss: 1.43761456013 Epoch: 35 -> Test Accuracy: 41.89 [36, 60] loss: 1.444 [36, 120] loss: 1.472 [36, 180] loss: 1.485 [36, 240] loss: 1.449 [36, 300] loss: 1.483 [36, 360] loss: 1.474 Epoch: 36 -> Loss: 1.41081655025 Epoch: 36 -> Test Accuracy: 41.5 [37, 60] loss: 1.456 [37, 120] loss: 1.457 [37, 180] loss: 1.454 [37, 240] loss: 1.480 [37, 300] loss: 1.475 [37, 360] loss: 1.473 Epoch: 37 -> Loss: 1.52845048904 Epoch: 37 -> Test Accuracy: 41.66 [38, 60] loss: 1.476 [38, 120] loss: 1.462 [38, 180] loss: 1.455 [38, 240] loss: 1.469 [38, 300] loss: 1.474 [38, 360] loss: 1.466 Epoch: 38 -> Loss: 1.4974886179 Epoch: 38 -> Test Accuracy: 41.39 [39, 60] loss: 1.481 [39, 120] loss: 1.476 [39, 180] loss: 1.477 [39, 240] loss: 1.454 [39, 300] loss: 1.461 [39, 360] loss: 1.460 Epoch: 39 -> Loss: 1.44558370113 Epoch: 39 -> Test Accuracy: 41.45 [40, 60] loss: 1.446 [40, 120] loss: 1.489 [40, 180] loss: 1.469 [40, 240] loss: 1.489 [40, 300] loss: 1.458 [40, 360] loss: 1.477 Epoch: 40 -> Loss: 1.40977978706 Epoch: 40 -> Test Accuracy: 41.07 [41, 60] loss: 1.449 [41, 120] loss: 1.434 [41, 180] loss: 1.425 [41, 240] loss: 1.417 [41, 300] loss: 1.394 [41, 360] loss: 1.410 Epoch: 41 -> Loss: 1.45218515396 Epoch: 41 -> Test Accuracy: 43.15 [42, 60] loss: 1.403 [42, 120] loss: 1.414 [42, 180] loss: 1.391 [42, 240] loss: 1.378 [42, 300] loss: 1.397 [42, 360] loss: 1.415 Epoch: 42 -> Loss: 1.45976185799 Epoch: 42 -> Test Accuracy: 43.57 [43, 60] loss: 1.410 [43, 120] loss: 1.389 [43, 180] loss: 1.407 [43, 240] loss: 1.369 [43, 300] loss: 1.384 [43, 360] loss: 1.405 Epoch: 43 -> Loss: 1.63363111019 Epoch: 43 -> Test Accuracy: 43.79 [44, 60] loss: 1.381 [44, 120] loss: 1.387 [44, 180] loss: 1.398 [44, 240] loss: 1.375 [44, 300] loss: 1.376 [44, 360] loss: 1.389 Epoch: 44 -> Loss: 1.39831256866 Epoch: 44 -> Test Accuracy: 44.2 [45, 60] loss: 1.369 [45, 120] loss: 1.387 [45, 180] loss: 1.391 [45, 240] loss: 1.374 [45, 300] loss: 1.381 [45, 360] loss: 1.390 Epoch: 45 -> Loss: 1.3232729435 Epoch: 45 -> Test Accuracy: 44.02 [46, 60] loss: 1.393 [46, 120] loss: 1.376 [46, 180] loss: 1.354 [46, 240] loss: 1.355 [46, 300] loss: 1.358 [46, 360] loss: 1.381 Epoch: 46 -> Loss: 1.37403297424 Epoch: 46 -> Test Accuracy: 44.28 [47, 60] loss: 1.339 [47, 120] loss: 1.371 [47, 180] loss: 1.383 [47, 240] loss: 1.374 [47, 300] loss: 1.355 [47, 360] loss: 1.366 Epoch: 47 -> Loss: 1.34893035889 Epoch: 47 -> Test Accuracy: 44.27 [48, 60] loss: 1.362 [48, 120] loss: 1.363 [48, 180] loss: 1.346 [48, 240] loss: 1.358 [48, 300] loss: 1.377 [48, 360] loss: 1.357 Epoch: 48 -> Loss: 1.38947319984 Epoch: 48 -> Test Accuracy: 44.51 [49, 60] loss: 1.352 [49, 120] loss: 1.351 [49, 180] loss: 1.360 [49, 240] loss: 1.366 [49, 300] loss: 1.353 [49, 360] loss: 1.356 Epoch: 49 -> Loss: 1.32876336575 Epoch: 49 -> Test Accuracy: 44.48 [50, 60] loss: 1.367 [50, 120] loss: 1.354 [50, 180] loss: 1.347 [50, 240] loss: 1.356 [50, 300] loss: 1.348 [50, 360] loss: 1.357 Epoch: 50 -> Loss: 1.3062171936 Epoch: 50 -> Test Accuracy: 44.77 [51, 60] loss: 1.327 [51, 120] loss: 1.362 [51, 180] loss: 1.375 [51, 240] loss: 1.357 [51, 300] loss: 1.330 [51, 360] loss: 1.359 Epoch: 51 -> Loss: 1.44059967995 Epoch: 51 -> Test Accuracy: 44.71 [52, 60] loss: 1.357 [52, 120] loss: 1.360 [52, 180] loss: 1.362 [52, 240] loss: 1.354 [52, 300] loss: 1.346 [52, 360] loss: 1.337 Epoch: 52 -> Loss: 1.40836787224 Epoch: 52 -> Test Accuracy: 44.7 [53, 60] loss: 1.354 [53, 120] loss: 1.364 [53, 180] loss: 1.354 [53, 240] loss: 1.359 [53, 300] loss: 1.354 [53, 360] loss: 1.357 Epoch: 53 -> Loss: 1.20832812786 Epoch: 53 -> Test Accuracy: 44.63 [54, 60] loss: 1.341 [54, 120] loss: 1.363 [54, 180] loss: 1.367 [54, 240] loss: 1.335 [54, 300] loss: 1.376 [54, 360] loss: 1.353 Epoch: 54 -> Loss: 1.49905264378 Epoch: 54 -> Test Accuracy: 44.6 [55, 60] loss: 1.362 [55, 120] loss: 1.350 [55, 180] loss: 1.344 [55, 240] loss: 1.346 [55, 300] loss: 1.338 [55, 360] loss: 1.343 Epoch: 55 -> Loss: 1.25955271721 Epoch: 55 -> Test Accuracy: 44.48 [56, 60] loss: 1.336 [56, 120] loss: 1.350 [56, 180] loss: 1.351 [56, 240] loss: 1.334 [56, 300] loss: 1.364 [56, 360] loss: 1.343 Epoch: 56 -> Loss: 1.45857298374 Epoch: 56 -> Test Accuracy: 44.86 [57, 60] loss: 1.354 [57, 120] loss: 1.326 [57, 180] loss: 1.362 [57, 240] loss: 1.328 [57, 300] loss: 1.353 [57, 360] loss: 1.348 Epoch: 57 -> Loss: 1.294267416 Epoch: 57 -> Test Accuracy: 44.36 [58, 60] loss: 1.353 [58, 120] loss: 1.363 [58, 180] loss: 1.342 [58, 240] loss: 1.341 [58, 300] loss: 1.338 [58, 360] loss: 1.343 Epoch: 58 -> Loss: 1.40837979317 Epoch: 58 -> Test Accuracy: 44.59 [59, 60] loss: 1.326 [59, 120] loss: 1.362 [59, 180] loss: 1.368 [59, 240] loss: 1.354 [59, 300] loss: 1.335 [59, 360] loss: 1.337 Epoch: 59 -> Loss: 1.46184694767 Epoch: 59 -> Test Accuracy: 44.7 [60, 60] loss: 1.327 [60, 120] loss: 1.338 [60, 180] loss: 1.325 [60, 240] loss: 1.345 [60, 300] loss: 1.353 [60, 360] loss: 1.333 Epoch: 60 -> Loss: 1.27309501171 Epoch: 60 -> Test Accuracy: 44.68 [61, 60] loss: 1.345 [61, 120] loss: 1.354 [61, 180] loss: 1.333 [61, 240] loss: 1.336 [61, 300] loss: 1.343 [61, 360] loss: 1.350 Epoch: 61 -> Loss: 1.40847659111 Epoch: 61 -> Test Accuracy: 44.65 [62, 60] loss: 1.334 [62, 120] loss: 1.338 [62, 180] loss: 1.338 [62, 240] loss: 1.340 [62, 300] loss: 1.336 [62, 360] loss: 1.347 Epoch: 62 -> Loss: 1.25489187241 Epoch: 62 -> Test Accuracy: 44.45 [63, 60] loss: 1.331 [63, 120] loss: 1.334 [63, 180] loss: 1.355 [63, 240] loss: 1.337 [63, 300] loss: 1.338 [63, 360] loss: 1.357 Epoch: 63 -> Loss: 1.25219511986 Epoch: 63 -> Test Accuracy: 44.55 [64, 60] loss: 1.345 [64, 120] loss: 1.359 [64, 180] loss: 1.329 [64, 240] loss: 1.332 [64, 300] loss: 1.338 [64, 360] loss: 1.354 Epoch: 64 -> Loss: 1.47248375416 Epoch: 64 -> Test Accuracy: 44.68 [65, 60] loss: 1.342 [65, 120] loss: 1.332 [65, 180] loss: 1.337 [65, 240] loss: 1.340 [65, 300] loss: 1.349 [65, 360] loss: 1.336 Epoch: 65 -> Loss: 1.36289978027 Epoch: 65 -> Test Accuracy: 44.87 [66, 60] loss: 1.353 [66, 120] loss: 1.336 [66, 180] loss: 1.343 [66, 240] loss: 1.337 [66, 300] loss: 1.324 [66, 360] loss: 1.348 Epoch: 66 -> Loss: 1.60350251198 Epoch: 66 -> Test Accuracy: 44.7 [67, 60] loss: 1.330 [67, 120] loss: 1.325 [67, 180] loss: 1.337 [67, 240] loss: 1.351 [67, 300] loss: 1.349 [67, 360] loss: 1.338 Epoch: 67 -> Loss: 1.08013629913 Epoch: 67 -> Test Accuracy: 44.8 [68, 60] loss: 1.349 [68, 120] loss: 1.338 [68, 180] loss: 1.330 [68, 240] loss: 1.338 [68, 300] loss: 1.358 [68, 360] loss: 1.352 Epoch: 68 -> Loss: 1.21484231949 Epoch: 68 -> Test Accuracy: 44.84 [69, 60] loss: 1.334 [69, 120] loss: 1.323 [69, 180] loss: 1.327 [69, 240] loss: 1.343 [69, 300] loss: 1.362 [69, 360] loss: 1.336 Epoch: 69 -> Loss: 1.54353761673 Epoch: 69 -> Test Accuracy: 44.87 [70, 60] loss: 1.346 [70, 120] loss: 1.333 [70, 180] loss: 1.340 [70, 240] loss: 1.329 [70, 300] loss: 1.348 [70, 360] loss: 1.338 Epoch: 70 -> Loss: 1.63164269924 Epoch: 70 -> Test Accuracy: 44.96 [71, 60] loss: 1.327 [71, 120] loss: 1.321 [71, 180] loss: 1.351 [71, 240] loss: 1.343 [71, 300] loss: 1.327 [71, 360] loss: 1.335 Epoch: 71 -> Loss: 1.33832609653 Epoch: 71 -> Test Accuracy: 45.0 [72, 60] loss: 1.335 [72, 120] loss: 1.326 [72, 180] loss: 1.350 [72, 240] loss: 1.336 [72, 300] loss: 1.348 [72, 360] loss: 1.349 Epoch: 72 -> Loss: 1.33676230907 Epoch: 72 -> Test Accuracy: 44.95 [73, 60] loss: 1.319 [73, 120] loss: 1.336 [73, 180] loss: 1.323 [73, 240] loss: 1.340 [73, 300] loss: 1.354 [73, 360] loss: 1.338 Epoch: 73 -> Loss: 1.20308446884 Epoch: 73 -> Test Accuracy: 45.18 [74, 60] loss: 1.348 [74, 120] loss: 1.337 [74, 180] loss: 1.334 [74, 240] loss: 1.338 [74, 300] loss: 1.319 [74, 360] loss: 1.336 Epoch: 74 -> Loss: 1.53289163113 Epoch: 74 -> Test Accuracy: 45.07 [75, 60] loss: 1.324 [75, 120] loss: 1.339 [75, 180] loss: 1.311 [75, 240] loss: 1.335 [75, 300] loss: 1.335 [75, 360] loss: 1.336 Epoch: 75 -> Loss: 1.37990307808 Epoch: 75 -> Test Accuracy: 45.04 [76, 60] loss: 1.342 [76, 120] loss: 1.320 [76, 180] loss: 1.315 [76, 240] loss: 1.357 [76, 300] loss: 1.336 [76, 360] loss: 1.339 Epoch: 76 -> Loss: 1.41355466843 Epoch: 76 -> Test Accuracy: 45.14 [77, 60] loss: 1.324 [77, 120] loss: 1.345 [77, 180] loss: 1.351 [77, 240] loss: 1.342 [77, 300] loss: 1.324 [77, 360] loss: 1.348 Epoch: 77 -> Loss: 1.48257124424 Epoch: 77 -> Test Accuracy: 45.17 [78, 60] loss: 1.323 [78, 120] loss: 1.323 [78, 180] loss: 1.324 [78, 240] loss: 1.344 [78, 300] loss: 1.335 [78, 360] loss: 1.316 Epoch: 78 -> Loss: 1.22074568272 Epoch: 78 -> Test Accuracy: 45.05 [79, 60] loss: 1.334 [79, 120] loss: 1.333 [79, 180] loss: 1.317 [79, 240] loss: 1.337 [79, 300] loss: 1.329 [79, 360] loss: 1.327 Epoch: 79 -> Loss: 1.44437217712 Epoch: 79 -> Test Accuracy: 45.29 [80, 60] loss: 1.316 [80, 120] loss: 1.326 [80, 180] loss: 1.327 [80, 240] loss: 1.321 [80, 300] loss: 1.320 [80, 360] loss: 1.338 Epoch: 80 -> Loss: 1.48346507549 Epoch: 80 -> Test Accuracy: 45.06 [81, 60] loss: 1.320 [81, 120] loss: 1.331 [81, 180] loss: 1.337 [81, 240] loss: 1.328 [81, 300] loss: 1.341 [81, 360] loss: 1.341 Epoch: 81 -> Loss: 1.38029301167 Epoch: 81 -> Test Accuracy: 45.24 [82, 60] loss: 1.314 [82, 120] loss: 1.349 [82, 180] loss: 1.312 [82, 240] loss: 1.326 [82, 300] loss: 1.327 [82, 360] loss: 1.338 Epoch: 82 -> Loss: 1.34826076031 Epoch: 82 -> Test Accuracy: 45.06 [83, 60] loss: 1.334 [83, 120] loss: 1.324 [83, 180] loss: 1.322 [83, 240] loss: 1.319 [83, 300] loss: 1.339 [83, 360] loss: 1.329 Epoch: 83 -> Loss: 1.21827435493 Epoch: 83 -> Test Accuracy: 45.21 [84, 60] loss: 1.321 [84, 120] loss: 1.328 [84, 180] loss: 1.354 [84, 240] loss: 1.327 [84, 300] loss: 1.334 [84, 360] loss: 1.314 Epoch: 84 -> Loss: 1.53800415993 Epoch: 84 -> Test Accuracy: 45.16 [85, 60] loss: 1.304 [85, 120] loss: 1.347 [85, 180] loss: 1.329 [85, 240] loss: 1.346 [85, 300] loss: 1.316 [85, 360] loss: 1.320 Epoch: 85 -> Loss: 1.34503436089 Epoch: 85 -> Test Accuracy: 45.05 [86, 60] loss: 1.318 [86, 120] loss: 1.345 [86, 180] loss: 1.312 [86, 240] loss: 1.334 [86, 300] loss: 1.336 [86, 360] loss: 1.325 Epoch: 86 -> Loss: 1.40680634975 Epoch: 86 -> Test Accuracy: 44.88 [87, 60] loss: 1.318 [87, 120] loss: 1.325 [87, 180] loss: 1.330 [87, 240] loss: 1.317 [87, 300] loss: 1.333 [87, 360] loss: 1.335 Epoch: 87 -> Loss: 1.33049619198 Epoch: 87 -> Test Accuracy: 45.14 [88, 60] loss: 1.315 [88, 120] loss: 1.311 [88, 180] loss: 1.316 [88, 240] loss: 1.325 [88, 300] loss: 1.340 [88, 360] loss: 1.331 Epoch: 88 -> Loss: 1.29452824593 Epoch: 88 -> Test Accuracy: 45.17 [89, 60] loss: 1.313 [89, 120] loss: 1.321 [89, 180] loss: 1.332 [89, 240] loss: 1.310 [89, 300] loss: 1.322 [89, 360] loss: 1.340 Epoch: 89 -> Loss: 1.57597136497 Epoch: 89 -> Test Accuracy: 44.82 [90, 60] loss: 1.317 [90, 120] loss: 1.322 [90, 180] loss: 1.337 [90, 240] loss: 1.314 [90, 300] loss: 1.324 [90, 360] loss: 1.331 Epoch: 90 -> Loss: 1.14589643478 Epoch: 90 -> Test Accuracy: 44.7 [91, 60] loss: 1.328 [91, 120] loss: 1.324 [91, 180] loss: 1.322 [91, 240] loss: 1.333 [91, 300] loss: 1.318 [91, 360] loss: 1.328 Epoch: 91 -> Loss: 1.24610877037 Epoch: 91 -> Test Accuracy: 45.18 [92, 60] loss: 1.322 [92, 120] loss: 1.321 [92, 180] loss: 1.321 [92, 240] loss: 1.334 [92, 300] loss: 1.307 [92, 360] loss: 1.316 Epoch: 92 -> Loss: 1.27793669701 Epoch: 92 -> Test Accuracy: 44.78 [93, 60] loss: 1.318 [93, 120] loss: 1.353 [93, 180] loss: 1.320 [93, 240] loss: 1.315 [93, 300] loss: 1.333 [93, 360] loss: 1.341 Epoch: 93 -> Loss: 1.31564640999 Epoch: 93 -> Test Accuracy: 45.0 [94, 60] loss: 1.305 [94, 120] loss: 1.308 [94, 180] loss: 1.306 [94, 240] loss: 1.343 [94, 300] loss: 1.336 [94, 360] loss: 1.320 Epoch: 94 -> Loss: 1.31199002266 Epoch: 94 -> Test Accuracy: 44.77 [95, 60] loss: 1.313 [95, 120] loss: 1.334 [95, 180] loss: 1.318 [95, 240] loss: 1.318 [95, 300] loss: 1.301 [95, 360] loss: 1.330 Epoch: 95 -> Loss: 1.41564643383 Epoch: 95 -> Test Accuracy: 44.84 [96, 60] loss: 1.318 [96, 120] loss: 1.314 [96, 180] loss: 1.329 [96, 240] loss: 1.326 [96, 300] loss: 1.322 [96, 360] loss: 1.315 Epoch: 96 -> Loss: 1.38825726509 Epoch: 96 -> Test Accuracy: 44.99 [97, 60] loss: 1.323 [97, 120] loss: 1.315 [97, 180] loss: 1.320 [97, 240] loss: 1.345 [97, 300] loss: 1.323 [97, 360] loss: 1.318 Epoch: 97 -> Loss: 1.39025008678 Epoch: 97 -> Test Accuracy: 44.8 [98, 60] loss: 1.313 [98, 120] loss: 1.302 [98, 180] loss: 1.326 [98, 240] loss: 1.338 [98, 300] loss: 1.325 [98, 360] loss: 1.321 Epoch: 98 -> Loss: 1.34853088856 Epoch: 98 -> Test Accuracy: 45.01 [99, 60] loss: 1.305 [99, 120] loss: 1.324 [99, 180] loss: 1.320 [99, 240] loss: 1.313 [99, 300] loss: 1.330 [99, 360] loss: 1.321 Epoch: 99 -> Loss: 1.39544785023 Epoch: 99 -> Test Accuracy: 45.17 [100, 60] loss: 1.313 [100, 120] loss: 1.312 [100, 180] loss: 1.313 [100, 240] loss: 1.315 [100, 300] loss: 1.314 [100, 360] loss: 1.305 Epoch: 100 -> Loss: 1.22449326515 Epoch: 100 -> Test Accuracy: 45.3 Finished Training
In [9]:
# train ConvClassifiers on feature map of net_3block
conv_block4_loss_log, _, conv_block4_test_accuracy_log, _, _ = tr.train_all_blocks(4, 10, [0.1, 0.02, 0.004, 0.0008],
[35, 70, 85, 100], 0.9, 5e-4, net_block4, criterion, trainloader, None, testloader, use_ConvClassifier=True)
[1, 60] loss: 1.372 [1, 120] loss: 1.074 [1, 180] loss: 0.946 [1, 240] loss: 0.909 [1, 300] loss: 0.884 [1, 360] loss: 0.838 Epoch: 1 -> Loss: 0.927259802818 Epoch: 1 -> Test Accuracy: 69.98 [2, 60] loss: 0.768 [2, 120] loss: 0.743 [2, 180] loss: 0.742 [2, 240] loss: 0.712 [2, 300] loss: 0.679 [2, 360] loss: 0.691 Epoch: 2 -> Loss: 0.617124915123 Epoch: 2 -> Test Accuracy: 74.04 [3, 60] loss: 0.652 [3, 120] loss: 0.655 [3, 180] loss: 0.645 [3, 240] loss: 0.647 [3, 300] loss: 0.624 [3, 360] loss: 0.621 Epoch: 3 -> Loss: 0.637347102165 Epoch: 3 -> Test Accuracy: 75.37 [4, 60] loss: 0.605 [4, 120] loss: 0.604 [4, 180] loss: 0.613 [4, 240] loss: 0.590 [4, 300] loss: 0.598 [4, 360] loss: 0.564 Epoch: 4 -> Loss: 0.387685835361 Epoch: 4 -> Test Accuracy: 76.34 [5, 60] loss: 0.557 [5, 120] loss: 0.569 [5, 180] loss: 0.550 [5, 240] loss: 0.555 [5, 300] loss: 0.573 [5, 360] loss: 0.573 Epoch: 5 -> Loss: 0.502129793167 Epoch: 5 -> Test Accuracy: 77.77 [6, 60] loss: 0.534 [6, 120] loss: 0.560 [6, 180] loss: 0.539 [6, 240] loss: 0.549 [6, 300] loss: 0.557 [6, 360] loss: 0.542 Epoch: 6 -> Loss: 0.361016571522 Epoch: 6 -> Test Accuracy: 78.28 [7, 60] loss: 0.539 [7, 120] loss: 0.514 [7, 180] loss: 0.537 [7, 240] loss: 0.529 [7, 300] loss: 0.514 [7, 360] loss: 0.538 Epoch: 7 -> Loss: 0.482561647892 Epoch: 7 -> Test Accuracy: 80.08 [8, 60] loss: 0.500 [8, 120] loss: 0.508 [8, 180] loss: 0.522 [8, 240] loss: 0.510 [8, 300] loss: 0.513 [8, 360] loss: 0.503 Epoch: 8 -> Loss: 0.55847465992 Epoch: 8 -> Test Accuracy: 80.35 [9, 60] loss: 0.480 [9, 120] loss: 0.501 [9, 180] loss: 0.497 [9, 240] loss: 0.501 [9, 300] loss: 0.522 [9, 360] loss: 0.518 Epoch: 9 -> Loss: 0.620239019394 Epoch: 9 -> Test Accuracy: 78.19 [10, 60] loss: 0.481 [10, 120] loss: 0.483 [10, 180] loss: 0.501 [10, 240] loss: 0.474 [10, 300] loss: 0.522 [10, 360] loss: 0.480 Epoch: 10 -> Loss: 0.398448884487 Epoch: 10 -> Test Accuracy: 79.7 [11, 60] loss: 0.488 [11, 120] loss: 0.488 [11, 180] loss: 0.479 [11, 240] loss: 0.488 [11, 300] loss: 0.497 [11, 360] loss: 0.489 Epoch: 11 -> Loss: 0.534637212753 Epoch: 11 -> Test Accuracy: 79.93 [12, 60] loss: 0.458 [12, 120] loss: 0.482 [12, 180] loss: 0.483 [12, 240] loss: 0.492 [12, 300] loss: 0.490 [12, 360] loss: 0.466 Epoch: 12 -> Loss: 0.386262238026 Epoch: 12 -> Test Accuracy: 80.51 [13, 60] loss: 0.434 [13, 120] loss: 0.467 [13, 180] loss: 0.458 [13, 240] loss: 0.472 [13, 300] loss: 0.493 [13, 360] loss: 0.497 Epoch: 13 -> Loss: 0.476347744465 Epoch: 13 -> Test Accuracy: 80.91 [14, 60] loss: 0.463 [14, 120] loss: 0.468 [14, 180] loss: 0.470 [14, 240] loss: 0.457 [14, 300] loss: 0.485 [14, 360] loss: 0.478 Epoch: 14 -> Loss: 0.516170620918 Epoch: 14 -> Test Accuracy: 80.41 [15, 60] loss: 0.459 [15, 120] loss: 0.452 [15, 180] loss: 0.461 [15, 240] loss: 0.465 [15, 300] loss: 0.464 [15, 360] loss: 0.487 Epoch: 15 -> Loss: 0.485541522503 Epoch: 15 -> Test Accuracy: 80.19 [16, 60] loss: 0.437 [16, 120] loss: 0.453 [16, 180] loss: 0.462 [16, 240] loss: 0.446 [16, 300] loss: 0.461 [16, 360] loss: 0.472 Epoch: 16 -> Loss: 0.498241096735 Epoch: 16 -> Test Accuracy: 80.32 [17, 60] loss: 0.437 [17, 120] loss: 0.427 [17, 180] loss: 0.472 [17, 240] loss: 0.462 [17, 300] loss: 0.460 [17, 360] loss: 0.471 Epoch: 17 -> Loss: 0.313609927893 Epoch: 17 -> Test Accuracy: 81.16 [18, 60] loss: 0.426 [18, 120] loss: 0.450 [18, 180] loss: 0.453 [18, 240] loss: 0.439 [18, 300] loss: 0.448 [18, 360] loss: 0.463 Epoch: 18 -> Loss: 0.437183380127 Epoch: 18 -> Test Accuracy: 81.39 [19, 60] loss: 0.436 [19, 120] loss: 0.452 [19, 180] loss: 0.450 [19, 240] loss: 0.438 [19, 300] loss: 0.461 [19, 360] loss: 0.452 Epoch: 19 -> Loss: 0.398821175098 Epoch: 19 -> Test Accuracy: 79.81 [20, 60] loss: 0.413 [20, 120] loss: 0.445 [20, 180] loss: 0.463 [20, 240] loss: 0.444 [20, 300] loss: 0.453 [20, 360] loss: 0.466 Epoch: 20 -> Loss: 0.49925032258 Epoch: 20 -> Test Accuracy: 79.81 [21, 60] loss: 0.415 [21, 120] loss: 0.427 [21, 180] loss: 0.439 [21, 240] loss: 0.468 [21, 300] loss: 0.454 [21, 360] loss: 0.472 Epoch: 21 -> Loss: 0.444497525692 Epoch: 21 -> Test Accuracy: 79.35 [22, 60] loss: 0.444 [22, 120] loss: 0.433 [22, 180] loss: 0.444 [22, 240] loss: 0.440 [22, 300] loss: 0.456 [22, 360] loss: 0.445 Epoch: 22 -> Loss: 0.283289670944 Epoch: 22 -> Test Accuracy: 80.5 [23, 60] loss: 0.403 [23, 120] loss: 0.427 [23, 180] loss: 0.452 [23, 240] loss: 0.448 [23, 300] loss: 0.450 [23, 360] loss: 0.450 Epoch: 23 -> Loss: 0.413048088551 Epoch: 23 -> Test Accuracy: 81.67 [24, 60] loss: 0.412 [24, 120] loss: 0.433 [24, 180] loss: 0.444 [24, 240] loss: 0.451 [24, 300] loss: 0.444 [24, 360] loss: 0.444 Epoch: 24 -> Loss: 0.227589562535 Epoch: 24 -> Test Accuracy: 81.36 [25, 60] loss: 0.420 [25, 120] loss: 0.437 [25, 180] loss: 0.449 [25, 240] loss: 0.428 [25, 300] loss: 0.453 [25, 360] loss: 0.442 Epoch: 25 -> Loss: 0.322181284428 Epoch: 25 -> Test Accuracy: 79.74 [26, 60] loss: 0.418 [26, 120] loss: 0.414 [26, 180] loss: 0.427 [26, 240] loss: 0.432 [26, 300] loss: 0.453 [26, 360] loss: 0.469 Epoch: 26 -> Loss: 0.476414352655 Epoch: 26 -> Test Accuracy: 80.22 [27, 60] loss: 0.411 [27, 120] loss: 0.422 [27, 180] loss: 0.429 [27, 240] loss: 0.439 [27, 300] loss: 0.448 [27, 360] loss: 0.436 Epoch: 27 -> Loss: 0.452017396688 Epoch: 27 -> Test Accuracy: 81.22 [28, 60] loss: 0.418 [28, 120] loss: 0.423 [28, 180] loss: 0.434 [28, 240] loss: 0.442 [28, 300] loss: 0.426 [28, 360] loss: 0.448 Epoch: 28 -> Loss: 0.456640064716 Epoch: 28 -> Test Accuracy: 81.29 [29, 60] loss: 0.402 [29, 120] loss: 0.421 [29, 180] loss: 0.441 [29, 240] loss: 0.442 [29, 300] loss: 0.449 [29, 360] loss: 0.451 Epoch: 29 -> Loss: 0.445857822895 Epoch: 29 -> Test Accuracy: 81.41 [30, 60] loss: 0.407 [30, 120] loss: 0.412 [30, 180] loss: 0.440 [30, 240] loss: 0.427 [30, 300] loss: 0.432 [30, 360] loss: 0.442 Epoch: 30 -> Loss: 0.485147058964 Epoch: 30 -> Test Accuracy: 80.7 [31, 60] loss: 0.410 [31, 120] loss: 0.425 [31, 180] loss: 0.422 [31, 240] loss: 0.427 [31, 300] loss: 0.426 [31, 360] loss: 0.452 Epoch: 31 -> Loss: 0.38826367259 Epoch: 31 -> Test Accuracy: 80.35 [32, 60] loss: 0.389 [32, 120] loss: 0.441 [32, 180] loss: 0.433 [32, 240] loss: 0.431 [32, 300] loss: 0.453 [32, 360] loss: 0.449 Epoch: 32 -> Loss: 0.395451396704 Epoch: 32 -> Test Accuracy: 81.43 [33, 60] loss: 0.402 [33, 120] loss: 0.422 [33, 180] loss: 0.441 [33, 240] loss: 0.442 [33, 300] loss: 0.435 [33, 360] loss: 0.428 Epoch: 33 -> Loss: 0.417356312275 Epoch: 33 -> Test Accuracy: 80.4 [34, 60] loss: 0.420 [34, 120] loss: 0.404 [34, 180] loss: 0.436 [34, 240] loss: 0.427 [34, 300] loss: 0.448 [34, 360] loss: 0.429 Epoch: 34 -> Loss: 0.32553473115 Epoch: 34 -> Test Accuracy: 79.84 [35, 60] loss: 0.410 [35, 120] loss: 0.412 [35, 180] loss: 0.413 [35, 240] loss: 0.443 [35, 300] loss: 0.438 [35, 360] loss: 0.447 Epoch: 35 -> Loss: 0.587553262711 Epoch: 35 -> Test Accuracy: 81.9 [36, 60] loss: 0.352 [36, 120] loss: 0.297 [36, 180] loss: 0.295 [36, 240] loss: 0.289 [36, 300] loss: 0.285 [36, 360] loss: 0.304 Epoch: 36 -> Loss: 0.188715487719 Epoch: 36 -> Test Accuracy: 85.13 [37, 60] loss: 0.270 [37, 120] loss: 0.279 [37, 180] loss: 0.266 [37, 240] loss: 0.274 [37, 300] loss: 0.269 [37, 360] loss: 0.256 Epoch: 37 -> Loss: 0.32158690691 Epoch: 37 -> Test Accuracy: 85.65 [38, 60] loss: 0.260 [38, 120] loss: 0.252 [38, 180] loss: 0.249 [38, 240] loss: 0.255 [38, 300] loss: 0.259 [38, 360] loss: 0.261 Epoch: 38 -> Loss: 0.306841343641 Epoch: 38 -> Test Accuracy: 84.64 [39, 60] loss: 0.235 [39, 120] loss: 0.259 [39, 180] loss: 0.241 [39, 240] loss: 0.245 [39, 300] loss: 0.255 [39, 360] loss: 0.255 Epoch: 39 -> Loss: 0.322778463364 Epoch: 39 -> Test Accuracy: 84.96 [40, 60] loss: 0.241 [40, 120] loss: 0.237 [40, 180] loss: 0.232 [40, 240] loss: 0.251 [40, 300] loss: 0.256 [40, 360] loss: 0.255 Epoch: 40 -> Loss: 0.195762485266 Epoch: 40 -> Test Accuracy: 85.03 [41, 60] loss: 0.225 [41, 120] loss: 0.240 [41, 180] loss: 0.241 [41, 240] loss: 0.235 [41, 300] loss: 0.256 [41, 360] loss: 0.243 Epoch: 41 -> Loss: 0.306314736605 Epoch: 41 -> Test Accuracy: 85.12 [42, 60] loss: 0.223 [42, 120] loss: 0.236 [42, 180] loss: 0.229 [42, 240] loss: 0.243 [42, 300] loss: 0.248 [42, 360] loss: 0.248 Epoch: 42 -> Loss: 0.209075495601 Epoch: 42 -> Test Accuracy: 84.72 [43, 60] loss: 0.228 [43, 120] loss: 0.228 [43, 180] loss: 0.220 [43, 240] loss: 0.245 [43, 300] loss: 0.242 [43, 360] loss: 0.252 Epoch: 43 -> Loss: 0.350984156132 Epoch: 43 -> Test Accuracy: 84.75 [44, 60] loss: 0.224 [44, 120] loss: 0.240 [44, 180] loss: 0.227 [44, 240] loss: 0.233 [44, 300] loss: 0.247 [44, 360] loss: 0.256 Epoch: 44 -> Loss: 0.243943408132 Epoch: 44 -> Test Accuracy: 85.28 [45, 60] loss: 0.230 [45, 120] loss: 0.229 [45, 180] loss: 0.231 [45, 240] loss: 0.235 [45, 300] loss: 0.243 [45, 360] loss: 0.252 Epoch: 45 -> Loss: 0.29595375061 Epoch: 45 -> Test Accuracy: 85.05 [46, 60] loss: 0.223 [46, 120] loss: 0.242 [46, 180] loss: 0.226 [46, 240] loss: 0.235 [46, 300] loss: 0.246 [46, 360] loss: 0.239 Epoch: 46 -> Loss: 0.233119890094 Epoch: 46 -> Test Accuracy: 84.48 [47, 60] loss: 0.214 [47, 120] loss: 0.230 [47, 180] loss: 0.236 [47, 240] loss: 0.254 [47, 300] loss: 0.236 [47, 360] loss: 0.244 Epoch: 47 -> Loss: 0.138906747103 Epoch: 47 -> Test Accuracy: 84.91 [48, 60] loss: 0.229 [48, 120] loss: 0.224 [48, 180] loss: 0.235 [48, 240] loss: 0.235 [48, 300] loss: 0.257 [48, 360] loss: 0.248 Epoch: 48 -> Loss: 0.282572805882 Epoch: 48 -> Test Accuracy: 84.55 [49, 60] loss: 0.232 [49, 120] loss: 0.236 [49, 180] loss: 0.245 [49, 240] loss: 0.231 [49, 300] loss: 0.229 [49, 360] loss: 0.249 Epoch: 49 -> Loss: 0.206251949072 Epoch: 49 -> Test Accuracy: 83.8 [50, 60] loss: 0.226 [50, 120] loss: 0.228 [50, 180] loss: 0.243 [50, 240] loss: 0.243 [50, 300] loss: 0.243 [50, 360] loss: 0.242 Epoch: 50 -> Loss: 0.20590980351 Epoch: 50 -> Test Accuracy: 84.09 [51, 60] loss: 0.220 [51, 120] loss: 0.233 [51, 180] loss: 0.225 [51, 240] loss: 0.243 [51, 300] loss: 0.237 [51, 360] loss: 0.255 Epoch: 51 -> Loss: 0.193142607808 Epoch: 51 -> Test Accuracy: 85.22 [52, 60] loss: 0.218 [52, 120] loss: 0.219 [52, 180] loss: 0.232 [52, 240] loss: 0.238 [52, 300] loss: 0.240 [52, 360] loss: 0.258 Epoch: 52 -> Loss: 0.203998044133 Epoch: 52 -> Test Accuracy: 84.16 [53, 60] loss: 0.222 [53, 120] loss: 0.239 [53, 180] loss: 0.225 [53, 240] loss: 0.232 [53, 300] loss: 0.251 [53, 360] loss: 0.237 Epoch: 53 -> Loss: 0.213248178363 Epoch: 53 -> Test Accuracy: 85.02 [54, 60] loss: 0.216 [54, 120] loss: 0.225 [54, 180] loss: 0.229 [54, 240] loss: 0.243 [54, 300] loss: 0.242 [54, 360] loss: 0.251 Epoch: 54 -> Loss: 0.288084477186 Epoch: 54 -> Test Accuracy: 84.64 [55, 60] loss: 0.215 [55, 120] loss: 0.218 [55, 180] loss: 0.232 [55, 240] loss: 0.251 [55, 300] loss: 0.239 [55, 360] loss: 0.234 Epoch: 55 -> Loss: 0.217760756612 Epoch: 55 -> Test Accuracy: 84.0 [56, 60] loss: 0.226 [56, 120] loss: 0.222 [56, 180] loss: 0.235 [56, 240] loss: 0.238 [56, 300] loss: 0.238 [56, 360] loss: 0.257 Epoch: 56 -> Loss: 0.300151914358 Epoch: 56 -> Test Accuracy: 84.23 [57, 60] loss: 0.224 [57, 120] loss: 0.233 [57, 180] loss: 0.236 [57, 240] loss: 0.227 [57, 300] loss: 0.243 [57, 360] loss: 0.235 Epoch: 57 -> Loss: 0.274967849255 Epoch: 57 -> Test Accuracy: 84.37 [58, 60] loss: 0.216 [58, 120] loss: 0.219 [58, 180] loss: 0.230 [58, 240] loss: 0.242 [58, 300] loss: 0.247 [58, 360] loss: 0.239 Epoch: 58 -> Loss: 0.264767020941 Epoch: 58 -> Test Accuracy: 84.05 [59, 60] loss: 0.235 [59, 120] loss: 0.225 [59, 180] loss: 0.217 [59, 240] loss: 0.227 [59, 300] loss: 0.235 [59, 360] loss: 0.249 Epoch: 59 -> Loss: 0.28347197175 Epoch: 59 -> Test Accuracy: 84.67 [60, 60] loss: 0.221 [60, 120] loss: 0.231 [60, 180] loss: 0.234 [60, 240] loss: 0.230 [60, 300] loss: 0.228 [60, 360] loss: 0.249 Epoch: 60 -> Loss: 0.229270026088 Epoch: 60 -> Test Accuracy: 83.87 [61, 60] loss: 0.219 [61, 120] loss: 0.223 [61, 180] loss: 0.237 [61, 240] loss: 0.231 [61, 300] loss: 0.237 [61, 360] loss: 0.258 Epoch: 61 -> Loss: 0.197796553373 Epoch: 61 -> Test Accuracy: 84.29 [62, 60] loss: 0.209 [62, 120] loss: 0.221 [62, 180] loss: 0.241 [62, 240] loss: 0.230 [62, 300] loss: 0.233 [62, 360] loss: 0.237 Epoch: 62 -> Loss: 0.221377894282 Epoch: 62 -> Test Accuracy: 84.52 [63, 60] loss: 0.204 [63, 120] loss: 0.220 [63, 180] loss: 0.234 [63, 240] loss: 0.240 [63, 300] loss: 0.236 [63, 360] loss: 0.249 Epoch: 63 -> Loss: 0.273113131523 Epoch: 63 -> Test Accuracy: 83.77 [64, 60] loss: 0.221 [64, 120] loss: 0.221 [64, 180] loss: 0.213 [64, 240] loss: 0.240 [64, 300] loss: 0.238 [64, 360] loss: 0.242 Epoch: 64 -> Loss: 0.244653537869 Epoch: 64 -> Test Accuracy: 84.29 [65, 60] loss: 0.218 [65, 120] loss: 0.221 [65, 180] loss: 0.233 [65, 240] loss: 0.225 [65, 300] loss: 0.231 [65, 360] loss: 0.226 Epoch: 65 -> Loss: 0.354703366756 Epoch: 65 -> Test Accuracy: 84.65 [66, 60] loss: 0.213 [66, 120] loss: 0.214 [66, 180] loss: 0.228 [66, 240] loss: 0.222 [66, 300] loss: 0.231 [66, 360] loss: 0.231 Epoch: 66 -> Loss: 0.25256896019 Epoch: 66 -> Test Accuracy: 84.73 [67, 60] loss: 0.213 [67, 120] loss: 0.214 [67, 180] loss: 0.224 [67, 240] loss: 0.231 [67, 300] loss: 0.224 [67, 360] loss: 0.233 Epoch: 67 -> Loss: 0.305030316114 Epoch: 67 -> Test Accuracy: 83.48 [68, 60] loss: 0.225 [68, 120] loss: 0.214 [68, 180] loss: 0.220 [68, 240] loss: 0.227 [68, 300] loss: 0.232 [68, 360] loss: 0.240 Epoch: 68 -> Loss: 0.171532616019 Epoch: 68 -> Test Accuracy: 84.13 [69, 60] loss: 0.218 [69, 120] loss: 0.226 [69, 180] loss: 0.227 [69, 240] loss: 0.216 [69, 300] loss: 0.229 [69, 360] loss: 0.239 Epoch: 69 -> Loss: 0.43925729394 Epoch: 69 -> Test Accuracy: 85.02 [70, 60] loss: 0.218 [70, 120] loss: 0.219 [70, 180] loss: 0.222 [70, 240] loss: 0.225 [70, 300] loss: 0.227 [70, 360] loss: 0.234 Epoch: 70 -> Loss: 0.268076896667 Epoch: 70 -> Test Accuracy: 83.92 [71, 60] loss: 0.182 [71, 120] loss: 0.163 [71, 180] loss: 0.154 [71, 240] loss: 0.160 [71, 300] loss: 0.154 [71, 360] loss: 0.148 Epoch: 71 -> Loss: 0.134753674269 Epoch: 71 -> Test Accuracy: 86.3 [72, 60] loss: 0.137 [72, 120] loss: 0.137 [72, 180] loss: 0.139 [72, 240] loss: 0.141 [72, 300] loss: 0.142 [72, 360] loss: 0.140 Epoch: 72 -> Loss: 0.126617431641 Epoch: 72 -> Test Accuracy: 86.51 [73, 60] loss: 0.131 [73, 120] loss: 0.132 [73, 180] loss: 0.133 [73, 240] loss: 0.128 [73, 300] loss: 0.136 [73, 360] loss: 0.129 Epoch: 73 -> Loss: 0.219107463956 Epoch: 73 -> Test Accuracy: 86.78 [74, 60] loss: 0.124 [74, 120] loss: 0.133 [74, 180] loss: 0.125 [74, 240] loss: 0.128 [74, 300] loss: 0.131 [74, 360] loss: 0.131 Epoch: 74 -> Loss: 0.130234628916 Epoch: 74 -> Test Accuracy: 86.51 [75, 60] loss: 0.120 [75, 120] loss: 0.123 [75, 180] loss: 0.124 [75, 240] loss: 0.125 [75, 300] loss: 0.126 [75, 360] loss: 0.123 Epoch: 75 -> Loss: 0.15600335598 Epoch: 75 -> Test Accuracy: 86.38 [76, 60] loss: 0.113 [76, 120] loss: 0.120 [76, 180] loss: 0.118 [76, 240] loss: 0.130 [76, 300] loss: 0.125 [76, 360] loss: 0.121 Epoch: 76 -> Loss: 0.161401793361 Epoch: 76 -> Test Accuracy: 86.8 [77, 60] loss: 0.116 [77, 120] loss: 0.113 [77, 180] loss: 0.116 [77, 240] loss: 0.120 [77, 300] loss: 0.131 [77, 360] loss: 0.116 Epoch: 77 -> Loss: 0.1019628793 Epoch: 77 -> Test Accuracy: 86.52 [78, 60] loss: 0.109 [78, 120] loss: 0.118 [78, 180] loss: 0.119 [78, 240] loss: 0.114 [78, 300] loss: 0.119 [78, 360] loss: 0.125 Epoch: 78 -> Loss: 0.163000136614 Epoch: 78 -> Test Accuracy: 86.54 [79, 60] loss: 0.114 [79, 120] loss: 0.109 [79, 180] loss: 0.120 [79, 240] loss: 0.120 [79, 300] loss: 0.117 [79, 360] loss: 0.116 Epoch: 79 -> Loss: 0.172471135855 Epoch: 79 -> Test Accuracy: 86.19 [80, 60] loss: 0.106 [80, 120] loss: 0.111 [80, 180] loss: 0.109 [80, 240] loss: 0.117 [80, 300] loss: 0.122 [80, 360] loss: 0.121 Epoch: 80 -> Loss: 0.145307347178 Epoch: 80 -> Test Accuracy: 85.97 [81, 60] loss: 0.108 [81, 120] loss: 0.105 [81, 180] loss: 0.107 [81, 240] loss: 0.119 [81, 300] loss: 0.109 [81, 360] loss: 0.120 Epoch: 81 -> Loss: 0.0871049165726 Epoch: 81 -> Test Accuracy: 86.18 [82, 60] loss: 0.096 [82, 120] loss: 0.114 [82, 180] loss: 0.108 [82, 240] loss: 0.115 [82, 300] loss: 0.110 [82, 360] loss: 0.115 Epoch: 82 -> Loss: 0.169233441353 Epoch: 82 -> Test Accuracy: 86.2 [83, 60] loss: 0.102 [83, 120] loss: 0.107 [83, 180] loss: 0.111 [83, 240] loss: 0.102 [83, 300] loss: 0.105 [83, 360] loss: 0.112 Epoch: 83 -> Loss: 0.085696414113 Epoch: 83 -> Test Accuracy: 86.61 [84, 60] loss: 0.102 [84, 120] loss: 0.107 [84, 180] loss: 0.109 [84, 240] loss: 0.112 [84, 300] loss: 0.105 [84, 360] loss: 0.109 Epoch: 84 -> Loss: 0.0708574280143 Epoch: 84 -> Test Accuracy: 86.27 [85, 60] loss: 0.105 [85, 120] loss: 0.100 [85, 180] loss: 0.100 [85, 240] loss: 0.106 [85, 300] loss: 0.114 [85, 360] loss: 0.104 Epoch: 85 -> Loss: 0.132660195231 Epoch: 85 -> Test Accuracy: 86.39 [86, 60] loss: 0.101 [86, 120] loss: 0.095 [86, 180] loss: 0.094 [86, 240] loss: 0.096 [86, 300] loss: 0.091 [86, 360] loss: 0.093 Epoch: 86 -> Loss: 0.175499588251 Epoch: 86 -> Test Accuracy: 86.56 [87, 60] loss: 0.088 [87, 120] loss: 0.096 [87, 180] loss: 0.090 [87, 240] loss: 0.089 [87, 300] loss: 0.093 [87, 360] loss: 0.090 Epoch: 87 -> Loss: 0.0678234323859 Epoch: 87 -> Test Accuracy: 86.6 [88, 60] loss: 0.088 [88, 120] loss: 0.090 [88, 180] loss: 0.097 [88, 240] loss: 0.093 [88, 300] loss: 0.093 [88, 360] loss: 0.088 Epoch: 88 -> Loss: 0.100074604154 Epoch: 88 -> Test Accuracy: 86.68 [89, 60] loss: 0.088 [89, 120] loss: 0.092 [89, 180] loss: 0.091 [89, 240] loss: 0.087 [89, 300] loss: 0.086 [89, 360] loss: 0.087 Epoch: 89 -> Loss: 0.0597632043064 Epoch: 89 -> Test Accuracy: 86.54 [90, 60] loss: 0.088 [90, 120] loss: 0.087 [90, 180] loss: 0.094 [90, 240] loss: 0.092 [90, 300] loss: 0.089 [90, 360] loss: 0.092 Epoch: 90 -> Loss: 0.102328419685 Epoch: 90 -> Test Accuracy: 86.37 [91, 60] loss: 0.082 [91, 120] loss: 0.083 [91, 180] loss: 0.088 [91, 240] loss: 0.088 [91, 300] loss: 0.095 [91, 360] loss: 0.085 Epoch: 91 -> Loss: 0.105287432671 Epoch: 91 -> Test Accuracy: 86.46 [92, 60] loss: 0.084 [92, 120] loss: 0.089 [92, 180] loss: 0.079 [92, 240] loss: 0.087 [92, 300] loss: 0.091 [92, 360] loss: 0.089 Epoch: 92 -> Loss: 0.0610167384148 Epoch: 92 -> Test Accuracy: 86.46 [93, 60] loss: 0.090 [93, 120] loss: 0.084 [93, 180] loss: 0.090 [93, 240] loss: 0.083 [93, 300] loss: 0.090 [93, 360] loss: 0.088 Epoch: 93 -> Loss: 0.122134879231 Epoch: 93 -> Test Accuracy: 86.46 [94, 60] loss: 0.089 [94, 120] loss: 0.086 [94, 180] loss: 0.084 [94, 240] loss: 0.090 [94, 300] loss: 0.088 [94, 360] loss: 0.086 Epoch: 94 -> Loss: 0.0983213037252 Epoch: 94 -> Test Accuracy: 86.52 [95, 60] loss: 0.083 [95, 120] loss: 0.085 [95, 180] loss: 0.084 [95, 240] loss: 0.088 [95, 300] loss: 0.089 [95, 360] loss: 0.088 Epoch: 95 -> Loss: 0.100874565542 Epoch: 95 -> Test Accuracy: 86.3 [96, 60] loss: 0.086 [96, 120] loss: 0.086 [96, 180] loss: 0.088 [96, 240] loss: 0.087 [96, 300] loss: 0.087 [96, 360] loss: 0.088 Epoch: 96 -> Loss: 0.0596638396382 Epoch: 96 -> Test Accuracy: 86.43 [97, 60] loss: 0.085 [97, 120] loss: 0.083 [97, 180] loss: 0.084 [97, 240] loss: 0.084 [97, 300] loss: 0.085 [97, 360] loss: 0.086 Epoch: 97 -> Loss: 0.104961775243 Epoch: 97 -> Test Accuracy: 86.52 [98, 60] loss: 0.083 [98, 120] loss: 0.086 [98, 180] loss: 0.088 [98, 240] loss: 0.088 [98, 300] loss: 0.086 [98, 360] loss: 0.082 Epoch: 98 -> Loss: 0.0625328570604 Epoch: 98 -> Test Accuracy: 86.51 [99, 60] loss: 0.084 [99, 120] loss: 0.085 [99, 180] loss: 0.081 [99, 240] loss: 0.085 [99, 300] loss: 0.086 [99, 360] loss: 0.092 Epoch: 99 -> Loss: 0.0963552966714 Epoch: 99 -> Test Accuracy: 86.59 [100, 60] loss: 0.084 [100, 120] loss: 0.082 [100, 180] loss: 0.082 [100, 240] loss: 0.085 [100, 300] loss: 0.087 [100, 360] loss: 0.090 Epoch: 100 -> Loss: 0.0826715677977 Epoch: 100 -> Test Accuracy: 86.58 Finished Training [1, 60] loss: 0.933 [1, 120] loss: 0.640 [1, 180] loss: 0.577 [1, 240] loss: 0.538 [1, 300] loss: 0.545 [1, 360] loss: 0.511 Epoch: 1 -> Loss: 0.605791211128 Epoch: 1 -> Test Accuracy: 81.44 [2, 60] loss: 0.435 [2, 120] loss: 0.434 [2, 180] loss: 0.447 [2, 240] loss: 0.456 [2, 300] loss: 0.436 [2, 360] loss: 0.420 Epoch: 2 -> Loss: 0.319448173046 Epoch: 2 -> Test Accuracy: 82.72 [3, 60] loss: 0.396 [3, 120] loss: 0.394 [3, 180] loss: 0.381 [3, 240] loss: 0.391 [3, 300] loss: 0.400 [3, 360] loss: 0.395 Epoch: 3 -> Loss: 0.357403695583 Epoch: 3 -> Test Accuracy: 83.92 [4, 60] loss: 0.348 [4, 120] loss: 0.357 [4, 180] loss: 0.371 [4, 240] loss: 0.362 [4, 300] loss: 0.357 [4, 360] loss: 0.366 Epoch: 4 -> Loss: 0.428539454937 Epoch: 4 -> Test Accuracy: 84.37 [5, 60] loss: 0.339 [5, 120] loss: 0.346 [5, 180] loss: 0.353 [5, 240] loss: 0.358 [5, 300] loss: 0.353 [5, 360] loss: 0.342 Epoch: 5 -> Loss: 0.266128063202 Epoch: 5 -> Test Accuracy: 84.48 [6, 60] loss: 0.319 [6, 120] loss: 0.326 [6, 180] loss: 0.343 [6, 240] loss: 0.344 [6, 300] loss: 0.343 [6, 360] loss: 0.336 Epoch: 6 -> Loss: 0.276509791613 Epoch: 6 -> Test Accuracy: 84.24 [7, 60] loss: 0.306 [7, 120] loss: 0.323 [7, 180] loss: 0.309 [7, 240] loss: 0.315 [7, 300] loss: 0.328 [7, 360] loss: 0.330 Epoch: 7 -> Loss: 0.277797162533 Epoch: 7 -> Test Accuracy: 84.96 [8, 60] loss: 0.295 [8, 120] loss: 0.332 [8, 180] loss: 0.302 [8, 240] loss: 0.314 [8, 300] loss: 0.319 [8, 360] loss: 0.321 Epoch: 8 -> Loss: 0.501798093319 Epoch: 8 -> Test Accuracy: 84.63 [9, 60] loss: 0.285 [9, 120] loss: 0.294 [9, 180] loss: 0.319 [9, 240] loss: 0.304 [9, 300] loss: 0.313 [9, 360] loss: 0.315 Epoch: 9 -> Loss: 0.305730760098 Epoch: 9 -> Test Accuracy: 85.47 [10, 60] loss: 0.279 [10, 120] loss: 0.302 [10, 180] loss: 0.294 [10, 240] loss: 0.309 [10, 300] loss: 0.313 [10, 360] loss: 0.308 Epoch: 10 -> Loss: 0.302999407053 Epoch: 10 -> Test Accuracy: 84.77 [11, 60] loss: 0.286 [11, 120] loss: 0.296 [11, 180] loss: 0.293 [11, 240] loss: 0.299 [11, 300] loss: 0.312 [11, 360] loss: 0.303 Epoch: 11 -> Loss: 0.191775470972 Epoch: 11 -> Test Accuracy: 85.41 [12, 60] loss: 0.265 [12, 120] loss: 0.265 [12, 180] loss: 0.298 [12, 240] loss: 0.298 [12, 300] loss: 0.308 [12, 360] loss: 0.305 Epoch: 12 -> Loss: 0.256272435188 Epoch: 12 -> Test Accuracy: 85.58 [13, 60] loss: 0.266 [13, 120] loss: 0.275 [13, 180] loss: 0.281 [13, 240] loss: 0.288 [13, 300] loss: 0.295 [13, 360] loss: 0.286 Epoch: 13 -> Loss: 0.412027359009 Epoch: 13 -> Test Accuracy: 85.37 [14, 60] loss: 0.262 [14, 120] loss: 0.269 [14, 180] loss: 0.281 [14, 240] loss: 0.288 [14, 300] loss: 0.303 [14, 360] loss: 0.286 Epoch: 14 -> Loss: 0.290439993143 Epoch: 14 -> Test Accuracy: 85.46 [15, 60] loss: 0.257 [15, 120] loss: 0.277 [15, 180] loss: 0.294 [15, 240] loss: 0.300 [15, 300] loss: 0.293 [15, 360] loss: 0.290 Epoch: 15 -> Loss: 0.317489922047 Epoch: 15 -> Test Accuracy: 85.22 [16, 60] loss: 0.271 [16, 120] loss: 0.267 [16, 180] loss: 0.266 [16, 240] loss: 0.283 [16, 300] loss: 0.301 [16, 360] loss: 0.295 Epoch: 16 -> Loss: 0.270550519228 Epoch: 16 -> Test Accuracy: 85.21 [17, 60] loss: 0.258 [17, 120] loss: 0.266 [17, 180] loss: 0.262 [17, 240] loss: 0.282 [17, 300] loss: 0.305 [17, 360] loss: 0.298 Epoch: 17 -> Loss: 0.317822188139 Epoch: 17 -> Test Accuracy: 85.85 [18, 60] loss: 0.249 [18, 120] loss: 0.273 [18, 180] loss: 0.274 [18, 240] loss: 0.295 [18, 300] loss: 0.282 [18, 360] loss: 0.287 Epoch: 18 -> Loss: 0.314208477736 Epoch: 18 -> Test Accuracy: 85.97 [19, 60] loss: 0.267 [19, 120] loss: 0.250 [19, 180] loss: 0.267 [19, 240] loss: 0.269 [19, 300] loss: 0.284 [19, 360] loss: 0.281 Epoch: 19 -> Loss: 0.19237627089 Epoch: 19 -> Test Accuracy: 85.6 [20, 60] loss: 0.254 [20, 120] loss: 0.264 [20, 180] loss: 0.265 [20, 240] loss: 0.275 [20, 300] loss: 0.292 [20, 360] loss: 0.284 Epoch: 20 -> Loss: 0.385714143515 Epoch: 20 -> Test Accuracy: 85.48 [21, 60] loss: 0.261 [21, 120] loss: 0.257 [21, 180] loss: 0.244 [21, 240] loss: 0.274 [21, 300] loss: 0.298 [21, 360] loss: 0.290 Epoch: 21 -> Loss: 0.356195032597 Epoch: 21 -> Test Accuracy: 86.2 [22, 60] loss: 0.254 [22, 120] loss: 0.251 [22, 180] loss: 0.262 [22, 240] loss: 0.272 [22, 300] loss: 0.293 [22, 360] loss: 0.265 Epoch: 22 -> Loss: 0.266731321812 Epoch: 22 -> Test Accuracy: 85.57 [23, 60] loss: 0.237 [23, 120] loss: 0.254 [23, 180] loss: 0.269 [23, 240] loss: 0.293 [23, 300] loss: 0.266 [23, 360] loss: 0.291 Epoch: 23 -> Loss: 0.296865284443 Epoch: 23 -> Test Accuracy: 85.48 [24, 60] loss: 0.242 [24, 120] loss: 0.267 [24, 180] loss: 0.269 [24, 240] loss: 0.271 [24, 300] loss: 0.289 [24, 360] loss: 0.283 Epoch: 24 -> Loss: 0.424262434244 Epoch: 24 -> Test Accuracy: 86.0 [25, 60] loss: 0.256 [25, 120] loss: 0.248 [25, 180] loss: 0.249 [25, 240] loss: 0.263 [25, 300] loss: 0.274 [25, 360] loss: 0.278 Epoch: 25 -> Loss: 0.484404802322 Epoch: 25 -> Test Accuracy: 85.9 [26, 60] loss: 0.254 [26, 120] loss: 0.251 [26, 180] loss: 0.257 [26, 240] loss: 0.275 [26, 300] loss: 0.282 [26, 360] loss: 0.276 Epoch: 26 -> Loss: 0.31919452548 Epoch: 26 -> Test Accuracy: 85.45 [27, 60] loss: 0.251 [27, 120] loss: 0.263 [27, 180] loss: 0.255 [27, 240] loss: 0.269 [27, 300] loss: 0.265 [27, 360] loss: 0.280 Epoch: 27 -> Loss: 0.288360774517 Epoch: 27 -> Test Accuracy: 85.97 [28, 60] loss: 0.228 [28, 120] loss: 0.259 [28, 180] loss: 0.267 [28, 240] loss: 0.271 [28, 300] loss: 0.283 [28, 360] loss: 0.278 Epoch: 28 -> Loss: 0.303036868572 Epoch: 28 -> Test Accuracy: 85.12 [29, 60] loss: 0.248 [29, 120] loss: 0.253 [29, 180] loss: 0.269 [29, 240] loss: 0.263 [29, 300] loss: 0.277 [29, 360] loss: 0.267 Epoch: 29 -> Loss: 0.328431010246 Epoch: 29 -> Test Accuracy: 85.1 [30, 60] loss: 0.250 [30, 120] loss: 0.230 [30, 180] loss: 0.274 [30, 240] loss: 0.278 [30, 300] loss: 0.272 [30, 360] loss: 0.285 Epoch: 30 -> Loss: 0.297442853451 Epoch: 30 -> Test Accuracy: 85.47 [31, 60] loss: 0.241 [31, 120] loss: 0.241 [31, 180] loss: 0.254 [31, 240] loss: 0.266 [31, 300] loss: 0.284 [31, 360] loss: 0.291 Epoch: 31 -> Loss: 0.275793671608 Epoch: 31 -> Test Accuracy: 85.57 [32, 60] loss: 0.234 [32, 120] loss: 0.257 [32, 180] loss: 0.252 [32, 240] loss: 0.264 [32, 300] loss: 0.278 [32, 360] loss: 0.286 Epoch: 32 -> Loss: 0.310254454613 Epoch: 32 -> Test Accuracy: 85.51 [33, 60] loss: 0.234 [33, 120] loss: 0.250 [33, 180] loss: 0.267 [33, 240] loss: 0.263 [33, 300] loss: 0.269 [33, 360] loss: 0.278 Epoch: 33 -> Loss: 0.424141407013 Epoch: 33 -> Test Accuracy: 85.2 [34, 60] loss: 0.228 [34, 120] loss: 0.240 [34, 180] loss: 0.257 [34, 240] loss: 0.272 [34, 300] loss: 0.271 [34, 360] loss: 0.275 Epoch: 34 -> Loss: 0.342402011156 Epoch: 34 -> Test Accuracy: 85.96 [35, 60] loss: 0.233 [35, 120] loss: 0.242 [35, 180] loss: 0.274 [35, 240] loss: 0.273 [35, 300] loss: 0.263 [35, 360] loss: 0.260 Epoch: 35 -> Loss: 0.506009280682 Epoch: 35 -> Test Accuracy: 85.96 [36, 60] loss: 0.202 [36, 120] loss: 0.186 [36, 180] loss: 0.174 [36, 240] loss: 0.174 [36, 300] loss: 0.154 [36, 360] loss: 0.170 Epoch: 36 -> Loss: 0.179405838251 Epoch: 36 -> Test Accuracy: 88.45 [37, 60] loss: 0.150 [37, 120] loss: 0.144 [37, 180] loss: 0.151 [37, 240] loss: 0.136 [37, 300] loss: 0.142 [37, 360] loss: 0.138 Epoch: 37 -> Loss: 0.0837014690042 Epoch: 37 -> Test Accuracy: 88.58 [38, 60] loss: 0.126 [38, 120] loss: 0.134 [38, 180] loss: 0.128 [38, 240] loss: 0.126 [38, 300] loss: 0.140 [38, 360] loss: 0.131 Epoch: 38 -> Loss: 0.134751647711 Epoch: 38 -> Test Accuracy: 88.53 [39, 60] loss: 0.123 [39, 120] loss: 0.124 [39, 180] loss: 0.112 [39, 240] loss: 0.126 [39, 300] loss: 0.123 [39, 360] loss: 0.125 Epoch: 39 -> Loss: 0.0824265927076 Epoch: 39 -> Test Accuracy: 88.36 [40, 60] loss: 0.114 [40, 120] loss: 0.109 [40, 180] loss: 0.116 [40, 240] loss: 0.115 [40, 300] loss: 0.122 [40, 360] loss: 0.128 Epoch: 40 -> Loss: 0.21214401722 Epoch: 40 -> Test Accuracy: 88.31 [41, 60] loss: 0.104 [41, 120] loss: 0.110 [41, 180] loss: 0.110 [41, 240] loss: 0.119 [41, 300] loss: 0.119 [41, 360] loss: 0.114 Epoch: 41 -> Loss: 0.263169586658 Epoch: 41 -> Test Accuracy: 88.04 [42, 60] loss: 0.104 [42, 120] loss: 0.105 [42, 180] loss: 0.111 [42, 240] loss: 0.117 [42, 300] loss: 0.106 [42, 360] loss: 0.111 Epoch: 42 -> Loss: 0.153781086206 Epoch: 42 -> Test Accuracy: 88.05 [43, 60] loss: 0.097 [43, 120] loss: 0.108 [43, 180] loss: 0.108 [43, 240] loss: 0.107 [43, 300] loss: 0.101 [43, 360] loss: 0.112 Epoch: 43 -> Loss: 0.156316131353 Epoch: 43 -> Test Accuracy: 87.71 [44, 60] loss: 0.097 [44, 120] loss: 0.106 [44, 180] loss: 0.107 [44, 240] loss: 0.108 [44, 300] loss: 0.111 [44, 360] loss: 0.106 Epoch: 44 -> Loss: 0.15221658349 Epoch: 44 -> Test Accuracy: 87.62 [45, 60] loss: 0.098 [45, 120] loss: 0.099 [45, 180] loss: 0.103 [45, 240] loss: 0.112 [45, 300] loss: 0.115 [45, 360] loss: 0.122 Epoch: 45 -> Loss: 0.0631106942892 Epoch: 45 -> Test Accuracy: 87.64 [46, 60] loss: 0.099 [46, 120] loss: 0.101 [46, 180] loss: 0.105 [46, 240] loss: 0.101 [46, 300] loss: 0.119 [46, 360] loss: 0.118 Epoch: 46 -> Loss: 0.109301820397 Epoch: 46 -> Test Accuracy: 87.78 [47, 60] loss: 0.098 [47, 120] loss: 0.096 [47, 180] loss: 0.111 [47, 240] loss: 0.109 [47, 300] loss: 0.112 [47, 360] loss: 0.118 Epoch: 47 -> Loss: 0.116154432297 Epoch: 47 -> Test Accuracy: 88.05 [48, 60] loss: 0.096 [48, 120] loss: 0.097 [48, 180] loss: 0.112 [48, 240] loss: 0.108 [48, 300] loss: 0.105 [48, 360] loss: 0.114 Epoch: 48 -> Loss: 0.0701258927584 Epoch: 48 -> Test Accuracy: 87.52 [49, 60] loss: 0.097 [49, 120] loss: 0.108 [49, 180] loss: 0.105 [49, 240] loss: 0.101 [49, 300] loss: 0.113 [49, 360] loss: 0.118 Epoch: 49 -> Loss: 0.0811708495021 Epoch: 49 -> Test Accuracy: 87.66 [50, 60] loss: 0.098 [50, 120] loss: 0.098 [50, 180] loss: 0.106 [50, 240] loss: 0.112 [50, 300] loss: 0.113 [50, 360] loss: 0.119 Epoch: 50 -> Loss: 0.165228754282 Epoch: 50 -> Test Accuracy: 87.59 [51, 60] loss: 0.097 [51, 120] loss: 0.096 [51, 180] loss: 0.105 [51, 240] loss: 0.111 [51, 300] loss: 0.117 [51, 360] loss: 0.106 Epoch: 51 -> Loss: 0.132050231099 Epoch: 51 -> Test Accuracy: 87.75 [52, 60] loss: 0.096 [52, 120] loss: 0.094 [52, 180] loss: 0.109 [52, 240] loss: 0.116 [52, 300] loss: 0.121 [52, 360] loss: 0.107 Epoch: 52 -> Loss: 0.0968780368567 Epoch: 52 -> Test Accuracy: 87.16 [53, 60] loss: 0.104 [53, 120] loss: 0.094 [53, 180] loss: 0.112 [53, 240] loss: 0.112 [53, 300] loss: 0.107 [53, 360] loss: 0.114 Epoch: 53 -> Loss: 0.0859619155526 Epoch: 53 -> Test Accuracy: 86.93 [54, 60] loss: 0.096 [54, 120] loss: 0.101 [54, 180] loss: 0.106 [54, 240] loss: 0.110 [54, 300] loss: 0.119 [54, 360] loss: 0.127 Epoch: 54 -> Loss: 0.139895915985 Epoch: 54 -> Test Accuracy: 87.57 [55, 60] loss: 0.099 [55, 120] loss: 0.096 [55, 180] loss: 0.111 [55, 240] loss: 0.113 [55, 300] loss: 0.106 [55, 360] loss: 0.111 Epoch: 55 -> Loss: 0.0857071876526 Epoch: 55 -> Test Accuracy: 87.28 [56, 60] loss: 0.105 [56, 120] loss: 0.114 [56, 180] loss: 0.110 [56, 240] loss: 0.109 [56, 300] loss: 0.116 [56, 360] loss: 0.113 Epoch: 56 -> Loss: 0.0610634274781 Epoch: 56 -> Test Accuracy: 87.17 [57, 60] loss: 0.100 [57, 120] loss: 0.105 [57, 180] loss: 0.105 [57, 240] loss: 0.108 [57, 300] loss: 0.116 [57, 360] loss: 0.111 Epoch: 57 -> Loss: 0.102903082967 Epoch: 57 -> Test Accuracy: 87.25 [58, 60] loss: 0.099 [58, 120] loss: 0.105 [58, 180] loss: 0.107 [58, 240] loss: 0.109 [58, 300] loss: 0.108 [58, 360] loss: 0.107 Epoch: 58 -> Loss: 0.158154025674 Epoch: 58 -> Test Accuracy: 86.9 [59, 60] loss: 0.094 [59, 120] loss: 0.106 [59, 180] loss: 0.102 [59, 240] loss: 0.105 [59, 300] loss: 0.125 [59, 360] loss: 0.110 Epoch: 59 -> Loss: 0.224472165108 Epoch: 59 -> Test Accuracy: 87.05 [60, 60] loss: 0.097 [60, 120] loss: 0.098 [60, 180] loss: 0.103 [60, 240] loss: 0.102 [60, 300] loss: 0.110 [60, 360] loss: 0.113 Epoch: 60 -> Loss: 0.148394331336 Epoch: 60 -> Test Accuracy: 87.09 [61, 60] loss: 0.106 [61, 120] loss: 0.103 [61, 180] loss: 0.098 [61, 240] loss: 0.116 [61, 300] loss: 0.118 [61, 360] loss: 0.111 Epoch: 61 -> Loss: 0.0785834789276 Epoch: 61 -> Test Accuracy: 86.8 [62, 60] loss: 0.110 [62, 120] loss: 0.095 [62, 180] loss: 0.100 [62, 240] loss: 0.099 [62, 300] loss: 0.107 [62, 360] loss: 0.116 Epoch: 62 -> Loss: 0.135267615318 Epoch: 62 -> Test Accuracy: 87.04 [63, 60] loss: 0.094 [63, 120] loss: 0.097 [63, 180] loss: 0.101 [63, 240] loss: 0.106 [63, 300] loss: 0.103 [63, 360] loss: 0.108 Epoch: 63 -> Loss: 0.0476409755647 Epoch: 63 -> Test Accuracy: 87.28 [64, 60] loss: 0.107 [64, 120] loss: 0.104 [64, 180] loss: 0.103 [64, 240] loss: 0.112 [64, 300] loss: 0.113 [64, 360] loss: 0.111 Epoch: 64 -> Loss: 0.140357613564 Epoch: 64 -> Test Accuracy: 87.35 [65, 60] loss: 0.090 [65, 120] loss: 0.092 [65, 180] loss: 0.119 [65, 240] loss: 0.109 [65, 300] loss: 0.115 [65, 360] loss: 0.110 Epoch: 65 -> Loss: 0.0916037410498 Epoch: 65 -> Test Accuracy: 87.69 [66, 60] loss: 0.094 [66, 120] loss: 0.096 [66, 180] loss: 0.102 [66, 240] loss: 0.102 [66, 300] loss: 0.105 [66, 360] loss: 0.114 Epoch: 66 -> Loss: 0.192954391241 Epoch: 66 -> Test Accuracy: 86.62 [67, 60] loss: 0.095 [67, 120] loss: 0.099 [67, 180] loss: 0.096 [67, 240] loss: 0.108 [67, 300] loss: 0.108 [67, 360] loss: 0.106 Epoch: 67 -> Loss: 0.109978698194 Epoch: 67 -> Test Accuracy: 87.2 [68, 60] loss: 0.094 [68, 120] loss: 0.094 [68, 180] loss: 0.097 [68, 240] loss: 0.105 [68, 300] loss: 0.109 [68, 360] loss: 0.108 Epoch: 68 -> Loss: 0.115408338606 Epoch: 68 -> Test Accuracy: 87.24 [69, 60] loss: 0.095 [69, 120] loss: 0.105 [69, 180] loss: 0.103 [69, 240] loss: 0.103 [69, 300] loss: 0.098 [69, 360] loss: 0.118 Epoch: 69 -> Loss: 0.173121526837 Epoch: 69 -> Test Accuracy: 87.27 [70, 60] loss: 0.090 [70, 120] loss: 0.090 [70, 180] loss: 0.103 [70, 240] loss: 0.105 [70, 300] loss: 0.103 [70, 360] loss: 0.114 Epoch: 70 -> Loss: 0.13461714983 Epoch: 70 -> Test Accuracy: 87.43 [71, 60] loss: 0.078 [71, 120] loss: 0.070 [71, 180] loss: 0.066 [71, 240] loss: 0.064 [71, 300] loss: 0.061 [71, 360] loss: 0.057 Epoch: 71 -> Loss: 0.0930706188083 Epoch: 71 -> Test Accuracy: 88.63 [72, 60] loss: 0.048 [72, 120] loss: 0.052 [72, 180] loss: 0.053 [72, 240] loss: 0.052 [72, 300] loss: 0.054 [72, 360] loss: 0.057 Epoch: 72 -> Loss: 0.0489098206162 Epoch: 72 -> Test Accuracy: 88.85 [73, 60] loss: 0.047 [73, 120] loss: 0.046 [73, 180] loss: 0.047 [73, 240] loss: 0.047 [73, 300] loss: 0.054 [73, 360] loss: 0.050 Epoch: 73 -> Loss: 0.0769120901823 Epoch: 73 -> Test Accuracy: 88.64 [74, 60] loss: 0.047 [74, 120] loss: 0.050 [74, 180] loss: 0.045 [74, 240] loss: 0.047 [74, 300] loss: 0.041 [74, 360] loss: 0.041 Epoch: 74 -> Loss: 0.069330945611 Epoch: 74 -> Test Accuracy: 88.6 [75, 60] loss: 0.042 [75, 120] loss: 0.040 [75, 180] loss: 0.045 [75, 240] loss: 0.046 [75, 300] loss: 0.042 [75, 360] loss: 0.042 Epoch: 75 -> Loss: 0.0492066368461 Epoch: 75 -> Test Accuracy: 89.07 [76, 60] loss: 0.037 [76, 120] loss: 0.037 [76, 180] loss: 0.046 [76, 240] loss: 0.042 [76, 300] loss: 0.038 [76, 360] loss: 0.038 Epoch: 76 -> Loss: 0.0620439946651 Epoch: 76 -> Test Accuracy: 88.78 [77, 60] loss: 0.037 [77, 120] loss: 0.038 [77, 180] loss: 0.038 [77, 240] loss: 0.040 [77, 300] loss: 0.041 [77, 360] loss: 0.040 Epoch: 77 -> Loss: 0.0635176822543 Epoch: 77 -> Test Accuracy: 88.66 [78, 60] loss: 0.037 [78, 120] loss: 0.035 [78, 180] loss: 0.035 [78, 240] loss: 0.036 [78, 300] loss: 0.039 [78, 360] loss: 0.038 Epoch: 78 -> Loss: 0.0321986787021 Epoch: 78 -> Test Accuracy: 88.69 [79, 60] loss: 0.037 [79, 120] loss: 0.037 [79, 180] loss: 0.039 [79, 240] loss: 0.038 [79, 300] loss: 0.036 [79, 360] loss: 0.036 Epoch: 79 -> Loss: 0.0234310366213 Epoch: 79 -> Test Accuracy: 88.74 [80, 60] loss: 0.035 [80, 120] loss: 0.034 [80, 180] loss: 0.033 [80, 240] loss: 0.034 [80, 300] loss: 0.036 [80, 360] loss: 0.036 Epoch: 80 -> Loss: 0.0715058892965 Epoch: 80 -> Test Accuracy: 88.65 [81, 60] loss: 0.032 [81, 120] loss: 0.034 [81, 180] loss: 0.034 [81, 240] loss: 0.036 [81, 300] loss: 0.037 [81, 360] loss: 0.037 Epoch: 81 -> Loss: 0.0276502557099 Epoch: 81 -> Test Accuracy: 88.6 [82, 60] loss: 0.032 [82, 120] loss: 0.031 [82, 180] loss: 0.035 [82, 240] loss: 0.033 [82, 300] loss: 0.031 [82, 360] loss: 0.037 Epoch: 82 -> Loss: 0.0351563431323 Epoch: 82 -> Test Accuracy: 88.44 [83, 60] loss: 0.031 [83, 120] loss: 0.032 [83, 180] loss: 0.033 [83, 240] loss: 0.033 [83, 300] loss: 0.036 [83, 360] loss: 0.036 Epoch: 83 -> Loss: 0.0192641858011 Epoch: 83 -> Test Accuracy: 88.77 [84, 60] loss: 0.030 [84, 120] loss: 0.032 [84, 180] loss: 0.032 [84, 240] loss: 0.033 [84, 300] loss: 0.033 [84, 360] loss: 0.032 Epoch: 84 -> Loss: 0.0821533873677 Epoch: 84 -> Test Accuracy: 88.63 [85, 60] loss: 0.032 [85, 120] loss: 0.032 [85, 180] loss: 0.031 [85, 240] loss: 0.029 [85, 300] loss: 0.034 [85, 360] loss: 0.032 Epoch: 85 -> Loss: 0.0377970822155 Epoch: 85 -> Test Accuracy: 88.53 [86, 60] loss: 0.029 [86, 120] loss: 0.033 [86, 180] loss: 0.028 [86, 240] loss: 0.026 [86, 300] loss: 0.030 [86, 360] loss: 0.026 Epoch: 86 -> Loss: 0.021526414901 Epoch: 86 -> Test Accuracy: 88.69 [87, 60] loss: 0.025 [87, 120] loss: 0.028 [87, 180] loss: 0.025 [87, 240] loss: 0.027 [87, 300] loss: 0.028 [87, 360] loss: 0.028 Epoch: 87 -> Loss: 0.014181887731 Epoch: 87 -> Test Accuracy: 88.59 [88, 60] loss: 0.027 [88, 120] loss: 0.026 [88, 180] loss: 0.028 [88, 240] loss: 0.028 [88, 300] loss: 0.029 [88, 360] loss: 0.029 Epoch: 88 -> Loss: 0.0407947227359 Epoch: 88 -> Test Accuracy: 88.67 [89, 60] loss: 0.025 [89, 120] loss: 0.026 [89, 180] loss: 0.028 [89, 240] loss: 0.027 [89, 300] loss: 0.026 [89, 360] loss: 0.028 Epoch: 89 -> Loss: 0.0337344445288 Epoch: 89 -> Test Accuracy: 88.64 [90, 60] loss: 0.026 [90, 120] loss: 0.029 [90, 180] loss: 0.025 [90, 240] loss: 0.027 [90, 300] loss: 0.026 [90, 360] loss: 0.025 Epoch: 90 -> Loss: 0.0341849885881 Epoch: 90 -> Test Accuracy: 88.64 [91, 60] loss: 0.030 [91, 120] loss: 0.027 [91, 180] loss: 0.030 [91, 240] loss: 0.027 [91, 300] loss: 0.027 [91, 360] loss: 0.027 Epoch: 91 -> Loss: 0.0218796730042 Epoch: 91 -> Test Accuracy: 88.8 [92, 60] loss: 0.028 [92, 120] loss: 0.024 [92, 180] loss: 0.028 [92, 240] loss: 0.025 [92, 300] loss: 0.030 [92, 360] loss: 0.028 Epoch: 92 -> Loss: 0.0230861660093 Epoch: 92 -> Test Accuracy: 88.66 [93, 60] loss: 0.028 [93, 120] loss: 0.028 [93, 180] loss: 0.024 [93, 240] loss: 0.025 [93, 300] loss: 0.028 [93, 360] loss: 0.026 Epoch: 93 -> Loss: 0.0308908764273 Epoch: 93 -> Test Accuracy: 88.7 [94, 60] loss: 0.025 [94, 120] loss: 0.026 [94, 180] loss: 0.025 [94, 240] loss: 0.024 [94, 300] loss: 0.026 [94, 360] loss: 0.027 Epoch: 94 -> Loss: 0.0687903538346 Epoch: 94 -> Test Accuracy: 88.73 [95, 60] loss: 0.027 [95, 120] loss: 0.027 [95, 180] loss: 0.025 [95, 240] loss: 0.026 [95, 300] loss: 0.028 [95, 360] loss: 0.026 Epoch: 95 -> Loss: 0.0634961277246 Epoch: 95 -> Test Accuracy: 88.84 [96, 60] loss: 0.026 [96, 120] loss: 0.026 [96, 180] loss: 0.028 [96, 240] loss: 0.026 [96, 300] loss: 0.025 [96, 360] loss: 0.027 Epoch: 96 -> Loss: 0.087438300252 Epoch: 96 -> Test Accuracy: 88.66 [97, 60] loss: 0.026 [97, 120] loss: 0.028 [97, 180] loss: 0.025 [97, 240] loss: 0.025 [97, 300] loss: 0.025 [97, 360] loss: 0.026 Epoch: 97 -> Loss: 0.0222023427486 Epoch: 97 -> Test Accuracy: 88.68 [98, 60] loss: 0.025 [98, 120] loss: 0.025 [98, 180] loss: 0.025 [98, 240] loss: 0.029 [98, 300] loss: 0.027 [98, 360] loss: 0.027 Epoch: 98 -> Loss: 0.0354787521064 Epoch: 98 -> Test Accuracy: 88.61 [99, 60] loss: 0.026 [99, 120] loss: 0.025 [99, 180] loss: 0.023 [99, 240] loss: 0.025 [99, 300] loss: 0.024 [99, 360] loss: 0.028 Epoch: 99 -> Loss: 0.0437642261386 Epoch: 99 -> Test Accuracy: 88.76 [100, 60] loss: 0.025 [100, 120] loss: 0.025 [100, 180] loss: 0.025 [100, 240] loss: 0.025 [100, 300] loss: 0.026 [100, 360] loss: 0.023 Epoch: 100 -> Loss: 0.0281120426953 Epoch: 100 -> Test Accuracy: 88.83 Finished Training [1, 60] loss: 0.904 [1, 120] loss: 0.697 [1, 180] loss: 0.629 [1, 240] loss: 0.602 [1, 300] loss: 0.577 [1, 360] loss: 0.555 Epoch: 1 -> Loss: 0.638784229755 Epoch: 1 -> Test Accuracy: 77.46 [2, 60] loss: 0.518 [2, 120] loss: 0.521 [2, 180] loss: 0.529 [2, 240] loss: 0.522 [2, 300] loss: 0.513 [2, 360] loss: 0.522 Epoch: 2 -> Loss: 0.496109187603 Epoch: 2 -> Test Accuracy: 79.71 [3, 60] loss: 0.500 [3, 120] loss: 0.474 [3, 180] loss: 0.494 [3, 240] loss: 0.487 [3, 300] loss: 0.472 [3, 360] loss: 0.486 Epoch: 3 -> Loss: 0.407837331295 Epoch: 3 -> Test Accuracy: 79.51 [4, 60] loss: 0.453 [4, 120] loss: 0.473 [4, 180] loss: 0.449 [4, 240] loss: 0.465 [4, 300] loss: 0.471 [4, 360] loss: 0.471 Epoch: 4 -> Loss: 0.487863689661 Epoch: 4 -> Test Accuracy: 79.92 [5, 60] loss: 0.451 [5, 120] loss: 0.446 [5, 180] loss: 0.450 [5, 240] loss: 0.444 [5, 300] loss: 0.453 [5, 360] loss: 0.476 Epoch: 5 -> Loss: 0.385122567415 Epoch: 5 -> Test Accuracy: 80.83 [6, 60] loss: 0.430 [6, 120] loss: 0.432 [6, 180] loss: 0.461 [6, 240] loss: 0.438 [6, 300] loss: 0.448 [6, 360] loss: 0.455 Epoch: 6 -> Loss: 0.413381874561 Epoch: 6 -> Test Accuracy: 80.09 [7, 60] loss: 0.431 [7, 120] loss: 0.432 [7, 180] loss: 0.424 [7, 240] loss: 0.440 [7, 300] loss: 0.436 [7, 360] loss: 0.432 Epoch: 7 -> Loss: 0.489655166864 Epoch: 7 -> Test Accuracy: 80.93 [8, 60] loss: 0.421 [8, 120] loss: 0.420 [8, 180] loss: 0.408 [8, 240] loss: 0.439 [8, 300] loss: 0.446 [8, 360] loss: 0.442 Epoch: 8 -> Loss: 0.447660923004 Epoch: 8 -> Test Accuracy: 80.85 [9, 60] loss: 0.398 [9, 120] loss: 0.424 [9, 180] loss: 0.427 [9, 240] loss: 0.434 [9, 300] loss: 0.431 [9, 360] loss: 0.423 Epoch: 9 -> Loss: 0.618254363537 Epoch: 9 -> Test Accuracy: 80.68 [10, 60] loss: 0.419 [10, 120] loss: 0.414 [10, 180] loss: 0.413 [10, 240] loss: 0.408 [10, 300] loss: 0.422 [10, 360] loss: 0.424 Epoch: 10 -> Loss: 0.370585739613 Epoch: 10 -> Test Accuracy: 81.33 [11, 60] loss: 0.386 [11, 120] loss: 0.400 [11, 180] loss: 0.411 [11, 240] loss: 0.417 [11, 300] loss: 0.438 [11, 360] loss: 0.415 Epoch: 11 -> Loss: 0.421869128942 Epoch: 11 -> Test Accuracy: 80.73 [12, 60] loss: 0.393 [12, 120] loss: 0.393 [12, 180] loss: 0.405 [12, 240] loss: 0.434 [12, 300] loss: 0.418 [12, 360] loss: 0.412 Epoch: 12 -> Loss: 0.379435688257 Epoch: 12 -> Test Accuracy: 81.68 [13, 60] loss: 0.401 [13, 120] loss: 0.414 [13, 180] loss: 0.410 [13, 240] loss: 0.432 [13, 300] loss: 0.393 [13, 360] loss: 0.417 Epoch: 13 -> Loss: 0.507173001766 Epoch: 13 -> Test Accuracy: 80.76 [14, 60] loss: 0.401 [14, 120] loss: 0.390 [14, 180] loss: 0.406 [14, 240] loss: 0.404 [14, 300] loss: 0.403 [14, 360] loss: 0.412 Epoch: 14 -> Loss: 0.511099457741 Epoch: 14 -> Test Accuracy: 81.39 [15, 60] loss: 0.389 [15, 120] loss: 0.389 [15, 180] loss: 0.405 [15, 240] loss: 0.407 [15, 300] loss: 0.405 [15, 360] loss: 0.423 Epoch: 15 -> Loss: 0.389810353518 Epoch: 15 -> Test Accuracy: 81.74 [16, 60] loss: 0.362 [16, 120] loss: 0.416 [16, 180] loss: 0.393 [16, 240] loss: 0.407 [16, 300] loss: 0.401 [16, 360] loss: 0.405 Epoch: 16 -> Loss: 0.300886750221 Epoch: 16 -> Test Accuracy: 80.07 [17, 60] loss: 0.393 [17, 120] loss: 0.392 [17, 180] loss: 0.407 [17, 240] loss: 0.403 [17, 300] loss: 0.393 [17, 360] loss: 0.387 Epoch: 17 -> Loss: 0.41320976615 Epoch: 17 -> Test Accuracy: 81.36 [18, 60] loss: 0.379 [18, 120] loss: 0.399 [18, 180] loss: 0.410 [18, 240] loss: 0.397 [18, 300] loss: 0.398 [18, 360] loss: 0.398 Epoch: 18 -> Loss: 0.485069662333 Epoch: 18 -> Test Accuracy: 81.31 [19, 60] loss: 0.368 [19, 120] loss: 0.391 [19, 180] loss: 0.408 [19, 240] loss: 0.403 [19, 300] loss: 0.414 [19, 360] loss: 0.390 Epoch: 19 -> Loss: 0.283340185881 Epoch: 19 -> Test Accuracy: 80.93 [20, 60] loss: 0.372 [20, 120] loss: 0.394 [20, 180] loss: 0.390 [20, 240] loss: 0.387 [20, 300] loss: 0.398 [20, 360] loss: 0.410 Epoch: 20 -> Loss: 0.516863584518 Epoch: 20 -> Test Accuracy: 80.23 [21, 60] loss: 0.383 [21, 120] loss: 0.379 [21, 180] loss: 0.394 [21, 240] loss: 0.390 [21, 300] loss: 0.397 [21, 360] loss: 0.417 Epoch: 21 -> Loss: 0.2647113204 Epoch: 21 -> Test Accuracy: 81.77 [22, 60] loss: 0.384 [22, 120] loss: 0.395 [22, 180] loss: 0.396 [22, 240] loss: 0.407 [22, 300] loss: 0.391 [22, 360] loss: 0.370 Epoch: 22 -> Loss: 0.390468150377 Epoch: 22 -> Test Accuracy: 82.19 [23, 60] loss: 0.377 [23, 120] loss: 0.375 [23, 180] loss: 0.375 [23, 240] loss: 0.404 [23, 300] loss: 0.406 [23, 360] loss: 0.397 Epoch: 23 -> Loss: 0.513204038143 Epoch: 23 -> Test Accuracy: 80.95 [24, 60] loss: 0.369 [24, 120] loss: 0.394 [24, 180] loss: 0.393 [24, 240] loss: 0.400 [24, 300] loss: 0.386 [24, 360] loss: 0.408 Epoch: 24 -> Loss: 0.329729020596 Epoch: 24 -> Test Accuracy: 81.18 [25, 60] loss: 0.374 [25, 120] loss: 0.379 [25, 180] loss: 0.365 [25, 240] loss: 0.403 [25, 300] loss: 0.391 [25, 360] loss: 0.404 Epoch: 25 -> Loss: 0.560485541821 Epoch: 25 -> Test Accuracy: 81.52 [26, 60] loss: 0.372 [26, 120] loss: 0.399 [26, 180] loss: 0.384 [26, 240] loss: 0.377 [26, 300] loss: 0.383 [26, 360] loss: 0.392 Epoch: 26 -> Loss: 0.372234463692 Epoch: 26 -> Test Accuracy: 81.32 [27, 60] loss: 0.389 [27, 120] loss: 0.379 [27, 180] loss: 0.387 [27, 240] loss: 0.393 [27, 300] loss: 0.379 [27, 360] loss: 0.391 Epoch: 27 -> Loss: 0.481274425983 Epoch: 27 -> Test Accuracy: 81.36 [28, 60] loss: 0.381 [28, 120] loss: 0.361 [28, 180] loss: 0.394 [28, 240] loss: 0.401 [28, 300] loss: 0.394 [28, 360] loss: 0.388 Epoch: 28 -> Loss: 0.351651012897 Epoch: 28 -> Test Accuracy: 81.88 [29, 60] loss: 0.371 [29, 120] loss: 0.373 [29, 180] loss: 0.375 [29, 240] loss: 0.375 [29, 300] loss: 0.417 [29, 360] loss: 0.399 Epoch: 29 -> Loss: 0.550917267799 Epoch: 29 -> Test Accuracy: 81.18 [30, 60] loss: 0.356 [30, 120] loss: 0.377 [30, 180] loss: 0.388 [30, 240] loss: 0.384 [30, 300] loss: 0.416 [30, 360] loss: 0.398 Epoch: 30 -> Loss: 0.294458210468 Epoch: 30 -> Test Accuracy: 81.2 [31, 60] loss: 0.373 [31, 120] loss: 0.386 [31, 180] loss: 0.360 [31, 240] loss: 0.410 [31, 300] loss: 0.390 [31, 360] loss: 0.393 Epoch: 31 -> Loss: 0.384102076292 Epoch: 31 -> Test Accuracy: 81.13 [32, 60] loss: 0.376 [32, 120] loss: 0.369 [32, 180] loss: 0.389 [32, 240] loss: 0.399 [32, 300] loss: 0.379 [32, 360] loss: 0.399 Epoch: 32 -> Loss: 0.405167967081 Epoch: 32 -> Test Accuracy: 81.24 [33, 60] loss: 0.361 [33, 120] loss: 0.394 [33, 180] loss: 0.382 [33, 240] loss: 0.378 [33, 300] loss: 0.400 [33, 360] loss: 0.391 Epoch: 33 -> Loss: 0.305773496628 Epoch: 33 -> Test Accuracy: 81.49 [34, 60] loss: 0.354 [34, 120] loss: 0.383 [34, 180] loss: 0.378 [34, 240] loss: 0.388 [34, 300] loss: 0.400 [34, 360] loss: 0.405 Epoch: 34 -> Loss: 0.448006093502 Epoch: 34 -> Test Accuracy: 81.39 [35, 60] loss: 0.352 [35, 120] loss: 0.389 [35, 180] loss: 0.385 [35, 240] loss: 0.379 [35, 300] loss: 0.375 [35, 360] loss: 0.398 Epoch: 35 -> Loss: 0.533167719841 Epoch: 35 -> Test Accuracy: 81.25 [36, 60] loss: 0.333 [36, 120] loss: 0.300 [36, 180] loss: 0.312 [36, 240] loss: 0.284 [36, 300] loss: 0.293 [36, 360] loss: 0.284 Epoch: 36 -> Loss: 0.323698431253 Epoch: 36 -> Test Accuracy: 84.19 [37, 60] loss: 0.279 [37, 120] loss: 0.283 [37, 180] loss: 0.273 [37, 240] loss: 0.278 [37, 300] loss: 0.286 [37, 360] loss: 0.276 Epoch: 37 -> Loss: 0.271359354258 Epoch: 37 -> Test Accuracy: 84.06 [38, 60] loss: 0.259 [38, 120] loss: 0.271 [38, 180] loss: 0.272 [38, 240] loss: 0.264 [38, 300] loss: 0.270 [38, 360] loss: 0.271 Epoch: 38 -> Loss: 0.403861284256 Epoch: 38 -> Test Accuracy: 83.78 [39, 60] loss: 0.247 [39, 120] loss: 0.258 [39, 180] loss: 0.234 [39, 240] loss: 0.253 [39, 300] loss: 0.261 [39, 360] loss: 0.257 Epoch: 39 -> Loss: 0.294805347919 Epoch: 39 -> Test Accuracy: 83.54 [40, 60] loss: 0.244 [40, 120] loss: 0.254 [40, 180] loss: 0.255 [40, 240] loss: 0.252 [40, 300] loss: 0.253 [40, 360] loss: 0.269 Epoch: 40 -> Loss: 0.23317758739 Epoch: 40 -> Test Accuracy: 83.88 [41, 60] loss: 0.249 [41, 120] loss: 0.245 [41, 180] loss: 0.246 [41, 240] loss: 0.248 [41, 300] loss: 0.256 [41, 360] loss: 0.269 Epoch: 41 -> Loss: 0.188509583473 Epoch: 41 -> Test Accuracy: 84.17 [42, 60] loss: 0.247 [42, 120] loss: 0.253 [42, 180] loss: 0.248 [42, 240] loss: 0.245 [42, 300] loss: 0.253 [42, 360] loss: 0.256 Epoch: 42 -> Loss: 0.288965761662 Epoch: 42 -> Test Accuracy: 83.76 [43, 60] loss: 0.233 [43, 120] loss: 0.248 [43, 180] loss: 0.244 [43, 240] loss: 0.262 [43, 300] loss: 0.244 [43, 360] loss: 0.258 Epoch: 43 -> Loss: 0.280441105366 Epoch: 43 -> Test Accuracy: 83.36 [44, 60] loss: 0.236 [44, 120] loss: 0.248 [44, 180] loss: 0.248 [44, 240] loss: 0.249 [44, 300] loss: 0.254 [44, 360] loss: 0.239 Epoch: 44 -> Loss: 0.329467117786 Epoch: 44 -> Test Accuracy: 83.52 [45, 60] loss: 0.239 [45, 120] loss: 0.226 [45, 180] loss: 0.243 [45, 240] loss: 0.236 [45, 300] loss: 0.264 [45, 360] loss: 0.260 Epoch: 45 -> Loss: 0.202561542392 Epoch: 45 -> Test Accuracy: 83.42 [46, 60] loss: 0.235 [46, 120] loss: 0.233 [46, 180] loss: 0.244 [46, 240] loss: 0.243 [46, 300] loss: 0.254 [46, 360] loss: 0.254 Epoch: 46 -> Loss: 0.427877992392 Epoch: 46 -> Test Accuracy: 83.4 [47, 60] loss: 0.229 [47, 120] loss: 0.246 [47, 180] loss: 0.236 [47, 240] loss: 0.231 [47, 300] loss: 0.252 [47, 360] loss: 0.263 Epoch: 47 -> Loss: 0.337016016245 Epoch: 47 -> Test Accuracy: 83.13 [48, 60] loss: 0.228 [48, 120] loss: 0.223 [48, 180] loss: 0.239 [48, 240] loss: 0.262 [48, 300] loss: 0.242 [48, 360] loss: 0.252 Epoch: 48 -> Loss: 0.292108207941 Epoch: 48 -> Test Accuracy: 82.98 [49, 60] loss: 0.232 [49, 120] loss: 0.235 [49, 180] loss: 0.244 [49, 240] loss: 0.255 [49, 300] loss: 0.236 [49, 360] loss: 0.254 Epoch: 49 -> Loss: 0.238557979465 Epoch: 49 -> Test Accuracy: 83.07 [50, 60] loss: 0.237 [50, 120] loss: 0.234 [50, 180] loss: 0.242 [50, 240] loss: 0.237 [50, 300] loss: 0.238 [50, 360] loss: 0.246 Epoch: 50 -> Loss: 0.215908616781 Epoch: 50 -> Test Accuracy: 83.12 [51, 60] loss: 0.224 [51, 120] loss: 0.231 [51, 180] loss: 0.227 [51, 240] loss: 0.252 [51, 300] loss: 0.251 [51, 360] loss: 0.254 Epoch: 51 -> Loss: 0.225903347135 Epoch: 51 -> Test Accuracy: 83.26 [52, 60] loss: 0.240 [52, 120] loss: 0.236 [52, 180] loss: 0.251 [52, 240] loss: 0.240 [52, 300] loss: 0.254 [52, 360] loss: 0.255 Epoch: 52 -> Loss: 0.3186224401 Epoch: 52 -> Test Accuracy: 83.42 [53, 60] loss: 0.238 [53, 120] loss: 0.234 [53, 180] loss: 0.223 [53, 240] loss: 0.239 [53, 300] loss: 0.250 [53, 360] loss: 0.252 Epoch: 53 -> Loss: 0.212616443634 Epoch: 53 -> Test Accuracy: 83.59 [54, 60] loss: 0.214 [54, 120] loss: 0.242 [54, 180] loss: 0.236 [54, 240] loss: 0.240 [54, 300] loss: 0.238 [54, 360] loss: 0.250 Epoch: 54 -> Loss: 0.227939888835 Epoch: 54 -> Test Accuracy: 83.62 [55, 60] loss: 0.233 [55, 120] loss: 0.234 [55, 180] loss: 0.244 [55, 240] loss: 0.246 [55, 300] loss: 0.247 [55, 360] loss: 0.239 Epoch: 55 -> Loss: 0.247250676155 Epoch: 55 -> Test Accuracy: 82.97 [56, 60] loss: 0.221 [56, 120] loss: 0.232 [56, 180] loss: 0.229 [56, 240] loss: 0.231 [56, 300] loss: 0.258 [56, 360] loss: 0.242 Epoch: 56 -> Loss: 0.195344880223 Epoch: 56 -> Test Accuracy: 83.28 [57, 60] loss: 0.223 [57, 120] loss: 0.233 [57, 180] loss: 0.238 [57, 240] loss: 0.226 [57, 300] loss: 0.234 [57, 360] loss: 0.241 Epoch: 57 -> Loss: 0.224278539419 Epoch: 57 -> Test Accuracy: 83.27 [58, 60] loss: 0.222 [58, 120] loss: 0.241 [58, 180] loss: 0.245 [58, 240] loss: 0.232 [58, 300] loss: 0.232 [58, 360] loss: 0.252 Epoch: 58 -> Loss: 0.162653848529 Epoch: 58 -> Test Accuracy: 83.1 [59, 60] loss: 0.228 [59, 120] loss: 0.232 [59, 180] loss: 0.224 [59, 240] loss: 0.233 [59, 300] loss: 0.237 [59, 360] loss: 0.235 Epoch: 59 -> Loss: 0.427598625422 Epoch: 59 -> Test Accuracy: 83.11 [60, 60] loss: 0.222 [60, 120] loss: 0.219 [60, 180] loss: 0.222 [60, 240] loss: 0.238 [60, 300] loss: 0.229 [60, 360] loss: 0.247 Epoch: 60 -> Loss: 0.244497850537 Epoch: 60 -> Test Accuracy: 82.93 [61, 60] loss: 0.220 [61, 120] loss: 0.213 [61, 180] loss: 0.231 [61, 240] loss: 0.219 [61, 300] loss: 0.240 [61, 360] loss: 0.247 Epoch: 61 -> Loss: 0.191278129816 Epoch: 61 -> Test Accuracy: 83.73 [62, 60] loss: 0.226 [62, 120] loss: 0.228 [62, 180] loss: 0.228 [62, 240] loss: 0.240 [62, 300] loss: 0.230 [62, 360] loss: 0.238 Epoch: 62 -> Loss: 0.288082003593 Epoch: 62 -> Test Accuracy: 83.21 [63, 60] loss: 0.216 [63, 120] loss: 0.223 [63, 180] loss: 0.226 [63, 240] loss: 0.244 [63, 300] loss: 0.229 [63, 360] loss: 0.233 Epoch: 63 -> Loss: 0.312612712383 Epoch: 63 -> Test Accuracy: 83.28 [64, 60] loss: 0.216 [64, 120] loss: 0.229 [64, 180] loss: 0.226 [64, 240] loss: 0.221 [64, 300] loss: 0.252 [64, 360] loss: 0.234 Epoch: 64 -> Loss: 0.145518258214 Epoch: 64 -> Test Accuracy: 83.6 [65, 60] loss: 0.219 [65, 120] loss: 0.222 [65, 180] loss: 0.230 [65, 240] loss: 0.236 [65, 300] loss: 0.229 [65, 360] loss: 0.224 Epoch: 65 -> Loss: 0.250201255083 Epoch: 65 -> Test Accuracy: 83.16 [66, 60] loss: 0.227 [66, 120] loss: 0.209 [66, 180] loss: 0.228 [66, 240] loss: 0.235 [66, 300] loss: 0.230 [66, 360] loss: 0.230 Epoch: 66 -> Loss: 0.285742133856 Epoch: 66 -> Test Accuracy: 83.06 [67, 60] loss: 0.224 [67, 120] loss: 0.219 [67, 180] loss: 0.230 [67, 240] loss: 0.227 [67, 300] loss: 0.234 [67, 360] loss: 0.235 Epoch: 67 -> Loss: 0.169107034802 Epoch: 67 -> Test Accuracy: 83.19 [68, 60] loss: 0.214 [68, 120] loss: 0.224 [68, 180] loss: 0.236 [68, 240] loss: 0.237 [68, 300] loss: 0.234 [68, 360] loss: 0.230 Epoch: 68 -> Loss: 0.157448172569 Epoch: 68 -> Test Accuracy: 83.38 [69, 60] loss: 0.207 [69, 120] loss: 0.205 [69, 180] loss: 0.228 [69, 240] loss: 0.223 [69, 300] loss: 0.241 [69, 360] loss: 0.229 Epoch: 69 -> Loss: 0.102319121361 Epoch: 69 -> Test Accuracy: 82.27 [70, 60] loss: 0.206 [70, 120] loss: 0.219 [70, 180] loss: 0.229 [70, 240] loss: 0.229 [70, 300] loss: 0.241 [70, 360] loss: 0.238 Epoch: 70 -> Loss: 0.20478323102 Epoch: 70 -> Test Accuracy: 82.68 [71, 60] loss: 0.202 [71, 120] loss: 0.178 [71, 180] loss: 0.174 [71, 240] loss: 0.169 [71, 300] loss: 0.166 [71, 360] loss: 0.162 Epoch: 71 -> Loss: 0.15055783093 Epoch: 71 -> Test Accuracy: 84.06 [72, 60] loss: 0.154 [72, 120] loss: 0.168 [72, 180] loss: 0.160 [72, 240] loss: 0.163 [72, 300] loss: 0.163 [72, 360] loss: 0.158 Epoch: 72 -> Loss: 0.249161928892 Epoch: 72 -> Test Accuracy: 84.42 [73, 60] loss: 0.152 [73, 120] loss: 0.153 [73, 180] loss: 0.147 [73, 240] loss: 0.151 [73, 300] loss: 0.143 [73, 360] loss: 0.156 Epoch: 73 -> Loss: 0.0970067381859 Epoch: 73 -> Test Accuracy: 84.54 [74, 60] loss: 0.147 [74, 120] loss: 0.139 [74, 180] loss: 0.145 [74, 240] loss: 0.162 [74, 300] loss: 0.153 [74, 360] loss: 0.141 Epoch: 74 -> Loss: 0.22745513916 Epoch: 74 -> Test Accuracy: 84.55 [75, 60] loss: 0.147 [75, 120] loss: 0.139 [75, 180] loss: 0.143 [75, 240] loss: 0.146 [75, 300] loss: 0.143 [75, 360] loss: 0.148 Epoch: 75 -> Loss: 0.170565709472 Epoch: 75 -> Test Accuracy: 84.31 [76, 60] loss: 0.146 [76, 120] loss: 0.140 [76, 180] loss: 0.142 [76, 240] loss: 0.138 [76, 300] loss: 0.146 [76, 360] loss: 0.139 Epoch: 76 -> Loss: 0.113970793784 Epoch: 76 -> Test Accuracy: 84.28 [77, 60] loss: 0.134 [77, 120] loss: 0.134 [77, 180] loss: 0.143 [77, 240] loss: 0.142 [77, 300] loss: 0.133 [77, 360] loss: 0.141 Epoch: 77 -> Loss: 0.0753377526999 Epoch: 77 -> Test Accuracy: 84.55 [78, 60] loss: 0.138 [78, 120] loss: 0.137 [78, 180] loss: 0.131 [78, 240] loss: 0.133 [78, 300] loss: 0.135 [78, 360] loss: 0.136 Epoch: 78 -> Loss: 0.303553342819 Epoch: 78 -> Test Accuracy: 84.35 [79, 60] loss: 0.129 [79, 120] loss: 0.130 [79, 180] loss: 0.130 [79, 240] loss: 0.134 [79, 300] loss: 0.135 [79, 360] loss: 0.125 Epoch: 79 -> Loss: 0.205618694425 Epoch: 79 -> Test Accuracy: 83.98 [80, 60] loss: 0.131 [80, 120] loss: 0.123 [80, 180] loss: 0.132 [80, 240] loss: 0.124 [80, 300] loss: 0.131 [80, 360] loss: 0.138 Epoch: 80 -> Loss: 0.188370659947 Epoch: 80 -> Test Accuracy: 84.16 [81, 60] loss: 0.126 [81, 120] loss: 0.136 [81, 180] loss: 0.130 [81, 240] loss: 0.129 [81, 300] loss: 0.118 [81, 360] loss: 0.131 Epoch: 81 -> Loss: 0.0754401236773 Epoch: 81 -> Test Accuracy: 84.42 [82, 60] loss: 0.123 [82, 120] loss: 0.121 [82, 180] loss: 0.127 [82, 240] loss: 0.134 [82, 300] loss: 0.129 [82, 360] loss: 0.129 Epoch: 82 -> Loss: 0.121076270938 Epoch: 82 -> Test Accuracy: 84.02 [83, 60] loss: 0.124 [83, 120] loss: 0.120 [83, 180] loss: 0.121 [83, 240] loss: 0.133 [83, 300] loss: 0.131 [83, 360] loss: 0.130 Epoch: 83 -> Loss: 0.229584142566 Epoch: 83 -> Test Accuracy: 83.99 [84, 60] loss: 0.115 [84, 120] loss: 0.126 [84, 180] loss: 0.125 [84, 240] loss: 0.116 [84, 300] loss: 0.118 [84, 360] loss: 0.124 Epoch: 84 -> Loss: 0.100957512856 Epoch: 84 -> Test Accuracy: 84.02 [85, 60] loss: 0.119 [85, 120] loss: 0.120 [85, 180] loss: 0.122 [85, 240] loss: 0.127 [85, 300] loss: 0.135 [85, 360] loss: 0.122 Epoch: 85 -> Loss: 0.131534129381 Epoch: 85 -> Test Accuracy: 83.75 [86, 60] loss: 0.116 [86, 120] loss: 0.105 [86, 180] loss: 0.107 [86, 240] loss: 0.115 [86, 300] loss: 0.111 [86, 360] loss: 0.111 Epoch: 86 -> Loss: 0.100786723197 Epoch: 86 -> Test Accuracy: 84.13 [87, 60] loss: 0.100 [87, 120] loss: 0.110 [87, 180] loss: 0.108 [87, 240] loss: 0.112 [87, 300] loss: 0.107 [87, 360] loss: 0.105 Epoch: 87 -> Loss: 0.144306018949 Epoch: 87 -> Test Accuracy: 84.2 [88, 60] loss: 0.105 [88, 120] loss: 0.104 [88, 180] loss: 0.106 [88, 240] loss: 0.100 [88, 300] loss: 0.107 [88, 360] loss: 0.107 Epoch: 88 -> Loss: 0.0612943395972 Epoch: 88 -> Test Accuracy: 84.24 [89, 60] loss: 0.111 [89, 120] loss: 0.104 [89, 180] loss: 0.104 [89, 240] loss: 0.102 [89, 300] loss: 0.100 [89, 360] loss: 0.110 Epoch: 89 -> Loss: 0.154033973813 Epoch: 89 -> Test Accuracy: 84.23 [90, 60] loss: 0.098 [90, 120] loss: 0.102 [90, 180] loss: 0.105 [90, 240] loss: 0.105 [90, 300] loss: 0.108 [90, 360] loss: 0.100 Epoch: 90 -> Loss: 0.146867543459 Epoch: 90 -> Test Accuracy: 84.18 [91, 60] loss: 0.101 [91, 120] loss: 0.100 [91, 180] loss: 0.106 [91, 240] loss: 0.109 [91, 300] loss: 0.100 [91, 360] loss: 0.098 Epoch: 91 -> Loss: 0.0843537226319 Epoch: 91 -> Test Accuracy: 84.15 [92, 60] loss: 0.101 [92, 120] loss: 0.097 [92, 180] loss: 0.104 [92, 240] loss: 0.097 [92, 300] loss: 0.104 [92, 360] loss: 0.104 Epoch: 92 -> Loss: 0.10631134361 Epoch: 92 -> Test Accuracy: 84.08 [93, 60] loss: 0.108 [93, 120] loss: 0.098 [93, 180] loss: 0.104 [93, 240] loss: 0.098 [93, 300] loss: 0.102 [93, 360] loss: 0.103 Epoch: 93 -> Loss: 0.0924872383475 Epoch: 93 -> Test Accuracy: 84.11 [94, 60] loss: 0.101 [94, 120] loss: 0.100 [94, 180] loss: 0.100 [94, 240] loss: 0.102 [94, 300] loss: 0.104 [94, 360] loss: 0.102 Epoch: 94 -> Loss: 0.139544397593 Epoch: 94 -> Test Accuracy: 84.28 [95, 60] loss: 0.099 [95, 120] loss: 0.101 [95, 180] loss: 0.098 [95, 240] loss: 0.102 [95, 300] loss: 0.095 [95, 360] loss: 0.099 Epoch: 95 -> Loss: 0.10687482357 Epoch: 95 -> Test Accuracy: 84.29 [96, 60] loss: 0.097 [96, 120] loss: 0.098 [96, 180] loss: 0.099 [96, 240] loss: 0.103 [96, 300] loss: 0.107 [96, 360] loss: 0.099 Epoch: 96 -> Loss: 0.0943318530917 Epoch: 96 -> Test Accuracy: 84.13 [97, 60] loss: 0.098 [97, 120] loss: 0.097 [97, 180] loss: 0.100 [97, 240] loss: 0.100 [97, 300] loss: 0.102 [97, 360] loss: 0.107 Epoch: 97 -> Loss: 0.070753082633 Epoch: 97 -> Test Accuracy: 84.34 [98, 60] loss: 0.102 [98, 120] loss: 0.097 [98, 180] loss: 0.102 [98, 240] loss: 0.094 [98, 300] loss: 0.096 [98, 360] loss: 0.097 Epoch: 98 -> Loss: 0.124475643039 Epoch: 98 -> Test Accuracy: 84.35 [99, 60] loss: 0.098 [99, 120] loss: 0.092 [99, 180] loss: 0.097 [99, 240] loss: 0.100 [99, 300] loss: 0.105 [99, 360] loss: 0.095 Epoch: 99 -> Loss: 0.135437041521 Epoch: 99 -> Test Accuracy: 84.19 [100, 60] loss: 0.096 [100, 120] loss: 0.092 [100, 180] loss: 0.094 [100, 240] loss: 0.109 [100, 300] loss: 0.099 [100, 360] loss: 0.101 Epoch: 100 -> Loss: 0.0642497316003 Epoch: 100 -> Test Accuracy: 84.25 Finished Training [1, 60] loss: 2.078 [1, 120] loss: 1.888 [1, 180] loss: 1.832 [1, 240] loss: 1.765 [1, 300] loss: 1.739 [1, 360] loss: 1.707 Epoch: 1 -> Loss: 1.54911959171 Epoch: 1 -> Test Accuracy: 33.97 [2, 60] loss: 1.692 [2, 120] loss: 1.663 [2, 180] loss: 1.658 [2, 240] loss: 1.640 [2, 300] loss: 1.648 [2, 360] loss: 1.617 Epoch: 2 -> Loss: 1.53013265133 Epoch: 2 -> Test Accuracy: 36.23 [3, 60] loss: 1.605 [3, 120] loss: 1.612 [3, 180] loss: 1.589 [3, 240] loss: 1.597 [3, 300] loss: 1.578 [3, 360] loss: 1.587 Epoch: 3 -> Loss: 1.47031855583 Epoch: 3 -> Test Accuracy: 38.1 [4, 60] loss: 1.544 [4, 120] loss: 1.563 [4, 180] loss: 1.591 [4, 240] loss: 1.556 [4, 300] loss: 1.541 [4, 360] loss: 1.572 Epoch: 4 -> Loss: 1.65348505974 Epoch: 4 -> Test Accuracy: 38.78 [5, 60] loss: 1.549 [5, 120] loss: 1.554 [5, 180] loss: 1.534 [5, 240] loss: 1.538 [5, 300] loss: 1.542 [5, 360] loss: 1.542 Epoch: 5 -> Loss: 1.35668408871 Epoch: 5 -> Test Accuracy: 40.06 [6, 60] loss: 1.528 [6, 120] loss: 1.515 [6, 180] loss: 1.525 [6, 240] loss: 1.531 [6, 300] loss: 1.526 [6, 360] loss: 1.512 Epoch: 6 -> Loss: 1.47895634174 Epoch: 6 -> Test Accuracy: 39.17 [7, 60] loss: 1.527 [7, 120] loss: 1.500 [7, 180] loss: 1.509 [7, 240] loss: 1.519 [7, 300] loss: 1.486 [7, 360] loss: 1.522 Epoch: 7 -> Loss: 1.44632053375 Epoch: 7 -> Test Accuracy: 40.48 [8, 60] loss: 1.494 [8, 120] loss: 1.527 [8, 180] loss: 1.483 [8, 240] loss: 1.484 [8, 300] loss: 1.505 [8, 360] loss: 1.493 Epoch: 8 -> Loss: 1.57093656063 Epoch: 8 -> Test Accuracy: 41.27 [9, 60] loss: 1.490 [9, 120] loss: 1.496 [9, 180] loss: 1.491 [9, 240] loss: 1.465 [9, 300] loss: 1.497 [9, 360] loss: 1.488 Epoch: 9 -> Loss: 1.37853515148 Epoch: 9 -> Test Accuracy: 40.99 [10, 60] loss: 1.471 [10, 120] loss: 1.467 [10, 180] loss: 1.497 [10, 240] loss: 1.494 [10, 300] loss: 1.482 [10, 360] loss: 1.502 Epoch: 10 -> Loss: 1.41953253746 Epoch: 10 -> Test Accuracy: 39.73 [11, 60] loss: 1.498 [11, 120] loss: 1.468 [11, 180] loss: 1.479 [11, 240] loss: 1.478 [11, 300] loss: 1.500 [11, 360] loss: 1.478 Epoch: 11 -> Loss: 1.52441179752 Epoch: 11 -> Test Accuracy: 39.65 [12, 60] loss: 1.487 [12, 120] loss: 1.482 [12, 180] loss: 1.476 [12, 240] loss: 1.470 [12, 300] loss: 1.499 [12, 360] loss: 1.471 Epoch: 12 -> Loss: 1.48645138741 Epoch: 12 -> Test Accuracy: 40.45 [13, 60] loss: 1.463 [13, 120] loss: 1.470 [13, 180] loss: 1.474 [13, 240] loss: 1.462 [13, 300] loss: 1.461 [13, 360] loss: 1.486 Epoch: 13 -> Loss: 1.38837504387 Epoch: 13 -> Test Accuracy: 41.59 [14, 60] loss: 1.482 [14, 120] loss: 1.472 [14, 180] loss: 1.468 [14, 240] loss: 1.479 [14, 300] loss: 1.464 [14, 360] loss: 1.469 Epoch: 14 -> Loss: 1.35015308857 Epoch: 14 -> Test Accuracy: 41.25 [15, 60] loss: 1.440 [15, 120] loss: 1.471 [15, 180] loss: 1.472 [15, 240] loss: 1.470 [15, 300] loss: 1.455 [15, 360] loss: 1.469 Epoch: 15 -> Loss: 1.38466477394 Epoch: 15 -> Test Accuracy: 41.58 [16, 60] loss: 1.464 [16, 120] loss: 1.478 [16, 180] loss: 1.457 [16, 240] loss: 1.459 [16, 300] loss: 1.434 [16, 360] loss: 1.446 Epoch: 16 -> Loss: 1.63905179501 Epoch: 16 -> Test Accuracy: 41.82 [17, 60] loss: 1.464 [17, 120] loss: 1.460 [17, 180] loss: 1.462 [17, 240] loss: 1.451 [17, 300] loss: 1.453 [17, 360] loss: 1.461 Epoch: 17 -> Loss: 1.62019217014 Epoch: 17 -> Test Accuracy: 41.56 [18, 60] loss: 1.470 [18, 120] loss: 1.439 [18, 180] loss: 1.446 [18, 240] loss: 1.465 [18, 300] loss: 1.463 [18, 360] loss: 1.432 Epoch: 18 -> Loss: 1.51963615417 Epoch: 18 -> Test Accuracy: 41.77 [19, 60] loss: 1.456 [19, 120] loss: 1.456 [19, 180] loss: 1.453 [19, 240] loss: 1.433 [19, 300] loss: 1.474 [19, 360] loss: 1.429 Epoch: 19 -> Loss: 1.47823345661 Epoch: 19 -> Test Accuracy: 41.85 [20, 60] loss: 1.439 [20, 120] loss: 1.451 [20, 180] loss: 1.448 [20, 240] loss: 1.433 [20, 300] loss: 1.445 [20, 360] loss: 1.458 Epoch: 20 -> Loss: 1.52669274807 Epoch: 20 -> Test Accuracy: 42.79 [21, 60] loss: 1.464 [21, 120] loss: 1.454 [21, 180] loss: 1.454 [21, 240] loss: 1.435 [21, 300] loss: 1.423 [21, 360] loss: 1.436 Epoch: 21 -> Loss: 1.40968728065 Epoch: 21 -> Test Accuracy: 42.19 [22, 60] loss: 1.437 [22, 120] loss: 1.442 [22, 180] loss: 1.455 [22, 240] loss: 1.460 [22, 300] loss: 1.466 [22, 360] loss: 1.442 Epoch: 22 -> Loss: 1.40106499195 Epoch: 22 -> Test Accuracy: 42.37 [23, 60] loss: 1.442 [23, 120] loss: 1.451 [23, 180] loss: 1.451 [23, 240] loss: 1.453 [23, 300] loss: 1.448 [23, 360] loss: 1.447 Epoch: 23 -> Loss: 1.31810796261 Epoch: 23 -> Test Accuracy: 42.31 [24, 60] loss: 1.470 [24, 120] loss: 1.437 [24, 180] loss: 1.446 [24, 240] loss: 1.442 [24, 300] loss: 1.452 [24, 360] loss: 1.458 Epoch: 24 -> Loss: 1.42444729805 Epoch: 24 -> Test Accuracy: 42.04 [25, 60] loss: 1.452 [25, 120] loss: 1.431 [25, 180] loss: 1.427 [25, 240] loss: 1.443 [25, 300] loss: 1.435 [25, 360] loss: 1.453 Epoch: 25 -> Loss: 1.44548606873 Epoch: 25 -> Test Accuracy: 40.63 [26, 60] loss: 1.440 [26, 120] loss: 1.433 [26, 180] loss: 1.449 [26, 240] loss: 1.442 [26, 300] loss: 1.451 [26, 360] loss: 1.449 Epoch: 26 -> Loss: 1.42789113522 Epoch: 26 -> Test Accuracy: 39.72 [27, 60] loss: 1.444 [27, 120] loss: 1.433 [27, 180] loss: 1.447 [27, 240] loss: 1.450 [27, 300] loss: 1.445 [27, 360] loss: 1.439 Epoch: 27 -> Loss: 1.29797279835 Epoch: 27 -> Test Accuracy: 42.67 [28, 60] loss: 1.452 [28, 120] loss: 1.436 [28, 180] loss: 1.443 [28, 240] loss: 1.457 [28, 300] loss: 1.438 [28, 360] loss: 1.435 Epoch: 28 -> Loss: 1.34364151955 Epoch: 28 -> Test Accuracy: 42.38 [29, 60] loss: 1.440 [29, 120] loss: 1.450 [29, 180] loss: 1.427 [29, 240] loss: 1.448 [29, 300] loss: 1.449 [29, 360] loss: 1.436 Epoch: 29 -> Loss: 1.34085488319 Epoch: 29 -> Test Accuracy: 40.93 [30, 60] loss: 1.413 [30, 120] loss: 1.448 [30, 180] loss: 1.446 [30, 240] loss: 1.458 [30, 300] loss: 1.430 [30, 360] loss: 1.444 Epoch: 30 -> Loss: 1.22373712063 Epoch: 30 -> Test Accuracy: 41.05 [31, 60] loss: 1.471 [31, 120] loss: 1.432 [31, 180] loss: 1.433 [31, 240] loss: 1.435 [31, 300] loss: 1.441 [31, 360] loss: 1.447 Epoch: 31 -> Loss: 1.49939751625 Epoch: 31 -> Test Accuracy: 42.81 [32, 60] loss: 1.445 [32, 120] loss: 1.422 [32, 180] loss: 1.428 [32, 240] loss: 1.455 [32, 300] loss: 1.459 [32, 360] loss: 1.438 Epoch: 32 -> Loss: 1.27505433559 Epoch: 32 -> Test Accuracy: 42.34 [33, 60] loss: 1.420 [33, 120] loss: 1.430 [33, 180] loss: 1.451 [33, 240] loss: 1.440 [33, 300] loss: 1.441 [33, 360] loss: 1.432 Epoch: 33 -> Loss: 1.40621566772 Epoch: 33 -> Test Accuracy: 42.84 [34, 60] loss: 1.443 [34, 120] loss: 1.443 [34, 180] loss: 1.428 [34, 240] loss: 1.436 [34, 300] loss: 1.445 [34, 360] loss: 1.435 Epoch: 34 -> Loss: 1.59400904179 Epoch: 34 -> Test Accuracy: 43.24 [35, 60] loss: 1.419 [35, 120] loss: 1.445 [35, 180] loss: 1.447 [35, 240] loss: 1.453 [35, 300] loss: 1.419 [35, 360] loss: 1.442 Epoch: 35 -> Loss: 1.48838484287 Epoch: 35 -> Test Accuracy: 41.14 [36, 60] loss: 1.346 [36, 120] loss: 1.326 [36, 180] loss: 1.339 [36, 240] loss: 1.338 [36, 300] loss: 1.318 [36, 360] loss: 1.316 Epoch: 36 -> Loss: 1.19623494148 Epoch: 36 -> Test Accuracy: 46.66 [37, 60] loss: 1.321 [37, 120] loss: 1.319 [37, 180] loss: 1.294 [37, 240] loss: 1.305 [37, 300] loss: 1.298 [37, 360] loss: 1.305 Epoch: 37 -> Loss: 1.50831675529 Epoch: 37 -> Test Accuracy: 46.22 [38, 60] loss: 1.312 [38, 120] loss: 1.264 [38, 180] loss: 1.304 [38, 240] loss: 1.297 [38, 300] loss: 1.296 [38, 360] loss: 1.306 Epoch: 38 -> Loss: 1.43125188351 Epoch: 38 -> Test Accuracy: 46.48 [39, 60] loss: 1.302 [39, 120] loss: 1.288 [39, 180] loss: 1.285 [39, 240] loss: 1.290 [39, 300] loss: 1.305 [39, 360] loss: 1.319 Epoch: 39 -> Loss: 1.29466211796 Epoch: 39 -> Test Accuracy: 47.21 [40, 60] loss: 1.303 [40, 120] loss: 1.295 [40, 180] loss: 1.282 [40, 240] loss: 1.290 [40, 300] loss: 1.294 [40, 360] loss: 1.298 Epoch: 40 -> Loss: 1.35837340355 Epoch: 40 -> Test Accuracy: 47.26 [41, 60] loss: 1.294 [41, 120] loss: 1.278 [41, 180] loss: 1.316 [41, 240] loss: 1.281 [41, 300] loss: 1.290 [41, 360] loss: 1.282 Epoch: 41 -> Loss: 1.26080310345 Epoch: 41 -> Test Accuracy: 47.14 [42, 60] loss: 1.275 [42, 120] loss: 1.296 [42, 180] loss: 1.292 [42, 240] loss: 1.284 [42, 300] loss: 1.302 [42, 360] loss: 1.283 Epoch: 42 -> Loss: 1.25724124908 Epoch: 42 -> Test Accuracy: 46.73 [43, 60] loss: 1.286 [43, 120] loss: 1.299 [43, 180] loss: 1.291 [43, 240] loss: 1.283 [43, 300] loss: 1.287 [43, 360] loss: 1.285 Epoch: 43 -> Loss: 1.5090290308 Epoch: 43 -> Test Accuracy: 46.82 [44, 60] loss: 1.291 [44, 120] loss: 1.299 [44, 180] loss: 1.292 [44, 240] loss: 1.291 [44, 300] loss: 1.276 [44, 360] loss: 1.277 Epoch: 44 -> Loss: 1.28428435326 Epoch: 44 -> Test Accuracy: 47.33 [45, 60] loss: 1.296 [45, 120] loss: 1.284 [45, 180] loss: 1.284 [45, 240] loss: 1.287 [45, 300] loss: 1.299 [45, 360] loss: 1.277 Epoch: 45 -> Loss: 1.26648044586 Epoch: 45 -> Test Accuracy: 47.16 [46, 60] loss: 1.290 [46, 120] loss: 1.269 [46, 180] loss: 1.272 [46, 240] loss: 1.295 [46, 300] loss: 1.311 [46, 360] loss: 1.294 Epoch: 46 -> Loss: 1.4980442524 Epoch: 46 -> Test Accuracy: 47.08 [47, 60] loss: 1.280 [47, 120] loss: 1.282 [47, 180] loss: 1.312 [47, 240] loss: 1.274 [47, 300] loss: 1.272 [47, 360] loss: 1.308 Epoch: 47 -> Loss: 1.35201942921 Epoch: 47 -> Test Accuracy: 46.76 [48, 60] loss: 1.300 [48, 120] loss: 1.277 [48, 180] loss: 1.279 [48, 240] loss: 1.296 [48, 300] loss: 1.269 [48, 360] loss: 1.306 Epoch: 48 -> Loss: 1.312510252 Epoch: 48 -> Test Accuracy: 46.17 [49, 60] loss: 1.277 [49, 120] loss: 1.296 [49, 180] loss: 1.277 [49, 240] loss: 1.282 [49, 300] loss: 1.292 [49, 360] loss: 1.282 Epoch: 49 -> Loss: 1.38111102581 Epoch: 49 -> Test Accuracy: 47.61 [50, 60] loss: 1.273 [50, 120] loss: 1.285 [50, 180] loss: 1.288 [50, 240] loss: 1.300 [50, 300] loss: 1.279 [50, 360] loss: 1.286 Epoch: 50 -> Loss: 1.38592135906 Epoch: 50 -> Test Accuracy: 46.45 [51, 60] loss: 1.285 [51, 120] loss: 1.294 [51, 180] loss: 1.278 [51, 240] loss: 1.279 [51, 300] loss: 1.274 [51, 360] loss: 1.285 Epoch: 51 -> Loss: 1.33155751228 Epoch: 51 -> Test Accuracy: 46.81 [52, 60] loss: 1.274 [52, 120] loss: 1.285 [52, 180] loss: 1.291 [52, 240] loss: 1.305 [52, 300] loss: 1.284 [52, 360] loss: 1.281 Epoch: 52 -> Loss: 1.2719591856 Epoch: 52 -> Test Accuracy: 48.22 [53, 60] loss: 1.280 [53, 120] loss: 1.274 [53, 180] loss: 1.293 [53, 240] loss: 1.302 [53, 300] loss: 1.278 [53, 360] loss: 1.288 Epoch: 53 -> Loss: 1.25599789619 Epoch: 53 -> Test Accuracy: 47.8 [54, 60] loss: 1.308 [54, 120] loss: 1.283 [54, 180] loss: 1.294 [54, 240] loss: 1.278 [54, 300] loss: 1.274 [54, 360] loss: 1.286 Epoch: 54 -> Loss: 1.26343917847 Epoch: 54 -> Test Accuracy: 47.52 [55, 60] loss: 1.272 [55, 120] loss: 1.265 [55, 180] loss: 1.291 [55, 240] loss: 1.296 [55, 300] loss: 1.307 [55, 360] loss: 1.282 Epoch: 55 -> Loss: 1.24955439568 Epoch: 55 -> Test Accuracy: 47.03 [56, 60] loss: 1.287 [56, 120] loss: 1.286 [56, 180] loss: 1.286 [56, 240] loss: 1.282 [56, 300] loss: 1.275 [56, 360] loss: 1.288 Epoch: 56 -> Loss: 1.33808410168 Epoch: 56 -> Test Accuracy: 47.26 [57, 60] loss: 1.283 [57, 120] loss: 1.288 [57, 180] loss: 1.284 [57, 240] loss: 1.289 [57, 300] loss: 1.279 [57, 360] loss: 1.277 Epoch: 57 -> Loss: 1.42024540901 Epoch: 57 -> Test Accuracy: 46.89 [58, 60] loss: 1.262 [58, 120] loss: 1.282 [58, 180] loss: 1.284 [58, 240] loss: 1.280 [58, 300] loss: 1.292 [58, 360] loss: 1.311 Epoch: 58 -> Loss: 1.24104607105 Epoch: 58 -> Test Accuracy: 48.01 [59, 60] loss: 1.296 [59, 120] loss: 1.275 [59, 180] loss: 1.273 [59, 240] loss: 1.274 [59, 300] loss: 1.285 [59, 360] loss: 1.269 Epoch: 59 -> Loss: 1.39220154285 Epoch: 59 -> Test Accuracy: 47.22 [60, 60] loss: 1.291 [60, 120] loss: 1.288 [60, 180] loss: 1.297 [60, 240] loss: 1.274 [60, 300] loss: 1.278 [60, 360] loss: 1.281 Epoch: 60 -> Loss: 1.36200547218 Epoch: 60 -> Test Accuracy: 47.83 [61, 60] loss: 1.276 [61, 120] loss: 1.274 [61, 180] loss: 1.283 [61, 240] loss: 1.294 [61, 300] loss: 1.281 [61, 360] loss: 1.280 Epoch: 61 -> Loss: 1.38666749001 Epoch: 61 -> Test Accuracy: 47.69 [62, 60] loss: 1.257 [62, 120] loss: 1.298 [62, 180] loss: 1.288 [62, 240] loss: 1.279 [62, 300] loss: 1.297 [62, 360] loss: 1.287 Epoch: 62 -> Loss: 1.23714363575 Epoch: 62 -> Test Accuracy: 46.32 [63, 60] loss: 1.282 [63, 120] loss: 1.264 [63, 180] loss: 1.283 [63, 240] loss: 1.287 [63, 300] loss: 1.282 [63, 360] loss: 1.278 Epoch: 63 -> Loss: 1.41803729534 Epoch: 63 -> Test Accuracy: 47.59 [64, 60] loss: 1.267 [64, 120] loss: 1.281 [64, 180] loss: 1.285 [64, 240] loss: 1.276 [64, 300] loss: 1.283 [64, 360] loss: 1.279 Epoch: 64 -> Loss: 1.23836088181 Epoch: 64 -> Test Accuracy: 47.52 [65, 60] loss: 1.270 [65, 120] loss: 1.271 [65, 180] loss: 1.285 [65, 240] loss: 1.280 [65, 300] loss: 1.293 [65, 360] loss: 1.288 Epoch: 65 -> Loss: 1.40561938286 Epoch: 65 -> Test Accuracy: 46.57 [66, 60] loss: 1.276 [66, 120] loss: 1.266 [66, 180] loss: 1.275 [66, 240] loss: 1.268 [66, 300] loss: 1.269 [66, 360] loss: 1.304 Epoch: 66 -> Loss: 1.28430604935 Epoch: 66 -> Test Accuracy: 47.95 [67, 60] loss: 1.277 [67, 120] loss: 1.267 [67, 180] loss: 1.270 [67, 240] loss: 1.283 [67, 300] loss: 1.280 [67, 360] loss: 1.275 Epoch: 67 -> Loss: 1.29519677162 Epoch: 67 -> Test Accuracy: 47.15 [68, 60] loss: 1.275 [68, 120] loss: 1.279 [68, 180] loss: 1.289 [68, 240] loss: 1.279 [68, 300] loss: 1.282 [68, 360] loss: 1.288 Epoch: 68 -> Loss: 1.24993872643 Epoch: 68 -> Test Accuracy: 46.99 [69, 60] loss: 1.265 [69, 120] loss: 1.267 [69, 180] loss: 1.260 [69, 240] loss: 1.272 [69, 300] loss: 1.258 [69, 360] loss: 1.281 Epoch: 69 -> Loss: 1.31426775455 Epoch: 69 -> Test Accuracy: 48.25 [70, 60] loss: 1.282 [70, 120] loss: 1.254 [70, 180] loss: 1.268 [70, 240] loss: 1.288 [70, 300] loss: 1.273 [70, 360] loss: 1.280 Epoch: 70 -> Loss: 1.25288391113 Epoch: 70 -> Test Accuracy: 48.14 [71, 60] loss: 1.203 [71, 120] loss: 1.212 [71, 180] loss: 1.209 [71, 240] loss: 1.200 [71, 300] loss: 1.215 [71, 360] loss: 1.181 Epoch: 71 -> Loss: 0.996318817139 Epoch: 71 -> Test Accuracy: 50.5 [72, 60] loss: 1.186 [72, 120] loss: 1.198 [72, 180] loss: 1.187 [72, 240] loss: 1.189 [72, 300] loss: 1.175 [72, 360] loss: 1.185 Epoch: 72 -> Loss: 1.13805437088 Epoch: 72 -> Test Accuracy: 50.65 [73, 60] loss: 1.170 [73, 120] loss: 1.182 [73, 180] loss: 1.175 [73, 240] loss: 1.195 [73, 300] loss: 1.208 [73, 360] loss: 1.180 Epoch: 73 -> Loss: 1.21059799194 Epoch: 73 -> Test Accuracy: 51.22 [74, 60] loss: 1.161 [74, 120] loss: 1.167 [74, 180] loss: 1.153 [74, 240] loss: 1.182 [74, 300] loss: 1.168 [74, 360] loss: 1.169 Epoch: 74 -> Loss: 1.24865567684 Epoch: 74 -> Test Accuracy: 51.04 [75, 60] loss: 1.159 [75, 120] loss: 1.161 [75, 180] loss: 1.161 [75, 240] loss: 1.162 [75, 300] loss: 1.191 [75, 360] loss: 1.172 Epoch: 75 -> Loss: 1.15342116356 Epoch: 75 -> Test Accuracy: 51.09 [76, 60] loss: 1.170 [76, 120] loss: 1.179 [76, 180] loss: 1.179 [76, 240] loss: 1.183 [76, 300] loss: 1.169 [76, 360] loss: 1.170 Epoch: 76 -> Loss: 1.09218108654 Epoch: 76 -> Test Accuracy: 50.97 [77, 60] loss: 1.154 [77, 120] loss: 1.180 [77, 180] loss: 1.161 [77, 240] loss: 1.160 [77, 300] loss: 1.176 [77, 360] loss: 1.152 Epoch: 77 -> Loss: 1.00372922421 Epoch: 77 -> Test Accuracy: 51.79 [78, 60] loss: 1.146 [78, 120] loss: 1.149 [78, 180] loss: 1.189 [78, 240] loss: 1.163 [78, 300] loss: 1.161 [78, 360] loss: 1.149 Epoch: 78 -> Loss: 1.34538543224 Epoch: 78 -> Test Accuracy: 51.11 [79, 60] loss: 1.156 [79, 120] loss: 1.174 [79, 180] loss: 1.164 [79, 240] loss: 1.161 [79, 300] loss: 1.169 [79, 360] loss: 1.152 Epoch: 79 -> Loss: 1.16095805168 Epoch: 79 -> Test Accuracy: 51.2 [80, 60] loss: 1.176 [80, 120] loss: 1.150 [80, 180] loss: 1.147 [80, 240] loss: 1.178 [80, 300] loss: 1.164 [80, 360] loss: 1.161 Epoch: 80 -> Loss: 1.29308915138 Epoch: 80 -> Test Accuracy: 51.42 [81, 60] loss: 1.161 [81, 120] loss: 1.159 [81, 180] loss: 1.150 [81, 240] loss: 1.153 [81, 300] loss: 1.166 [81, 360] loss: 1.143 Epoch: 81 -> Loss: 1.29539585114 Epoch: 81 -> Test Accuracy: 51.43 [82, 60] loss: 1.137 [82, 120] loss: 1.164 [82, 180] loss: 1.144 [82, 240] loss: 1.188 [82, 300] loss: 1.161 [82, 360] loss: 1.145 Epoch: 82 -> Loss: 1.00073957443 Epoch: 82 -> Test Accuracy: 51.37 [83, 60] loss: 1.176 [83, 120] loss: 1.162 [83, 180] loss: 1.152 [83, 240] loss: 1.161 [83, 300] loss: 1.167 [83, 360] loss: 1.162 Epoch: 83 -> Loss: 1.1242620945 Epoch: 83 -> Test Accuracy: 52.61 [84, 60] loss: 1.164 [84, 120] loss: 1.163 [84, 180] loss: 1.157 [84, 240] loss: 1.173 [84, 300] loss: 1.143 [84, 360] loss: 1.169 Epoch: 84 -> Loss: 1.27708125114 Epoch: 84 -> Test Accuracy: 50.68 [85, 60] loss: 1.162 [85, 120] loss: 1.176 [85, 180] loss: 1.158 [85, 240] loss: 1.163 [85, 300] loss: 1.162 [85, 360] loss: 1.141 Epoch: 85 -> Loss: 1.06103205681 Epoch: 85 -> Test Accuracy: 51.77 [86, 60] loss: 1.135 [86, 120] loss: 1.119 [86, 180] loss: 1.135 [86, 240] loss: 1.115 [86, 300] loss: 1.122 [86, 360] loss: 1.138 Epoch: 86 -> Loss: 1.16500258446 Epoch: 86 -> Test Accuracy: 52.5 [87, 60] loss: 1.108 [87, 120] loss: 1.115 [87, 180] loss: 1.143 [87, 240] loss: 1.094 [87, 300] loss: 1.123 [87, 360] loss: 1.102 Epoch: 87 -> Loss: 1.16233706474 Epoch: 87 -> Test Accuracy: 52.94 [88, 60] loss: 1.117 [88, 120] loss: 1.119 [88, 180] loss: 1.106 [88, 240] loss: 1.121 [88, 300] loss: 1.105 [88, 360] loss: 1.122 Epoch: 88 -> Loss: 0.972873389721 Epoch: 88 -> Test Accuracy: 52.66 [89, 60] loss: 1.104 [89, 120] loss: 1.094 [89, 180] loss: 1.130 [89, 240] loss: 1.099 [89, 300] loss: 1.113 [89, 360] loss: 1.107 Epoch: 89 -> Loss: 1.1231405735 Epoch: 89 -> Test Accuracy: 53.01 [90, 60] loss: 1.101 [90, 120] loss: 1.120 [90, 180] loss: 1.116 [90, 240] loss: 1.121 [90, 300] loss: 1.114 [90, 360] loss: 1.110 Epoch: 90 -> Loss: 0.941604614258 Epoch: 90 -> Test Accuracy: 53.25 [91, 60] loss: 1.113 [91, 120] loss: 1.102 [91, 180] loss: 1.117 [91, 240] loss: 1.098 [91, 300] loss: 1.109 [91, 360] loss: 1.102 Epoch: 91 -> Loss: 1.06780660152 Epoch: 91 -> Test Accuracy: 52.86 [92, 60] loss: 1.101 [92, 120] loss: 1.123 [92, 180] loss: 1.095 [92, 240] loss: 1.115 [92, 300] loss: 1.097 [92, 360] loss: 1.109 Epoch: 92 -> Loss: 1.26290023327 Epoch: 92 -> Test Accuracy: 53.22 [93, 60] loss: 1.111 [93, 120] loss: 1.121 [93, 180] loss: 1.092 [93, 240] loss: 1.112 [93, 300] loss: 1.102 [93, 360] loss: 1.101 Epoch: 93 -> Loss: 1.0867921114 Epoch: 93 -> Test Accuracy: 53.11 [94, 60] loss: 1.104 [94, 120] loss: 1.115 [94, 180] loss: 1.083 [94, 240] loss: 1.112 [94, 300] loss: 1.108 [94, 360] loss: 1.111 Epoch: 94 -> Loss: 1.05639493465 Epoch: 94 -> Test Accuracy: 53.47 [95, 60] loss: 1.113 [95, 120] loss: 1.090 [95, 180] loss: 1.112 [95, 240] loss: 1.124 [95, 300] loss: 1.106 [95, 360] loss: 1.104 Epoch: 95 -> Loss: 1.07176852226 Epoch: 95 -> Test Accuracy: 52.88 [96, 60] loss: 1.099 [96, 120] loss: 1.113 [96, 180] loss: 1.115 [96, 240] loss: 1.088 [96, 300] loss: 1.105 [96, 360] loss: 1.111 Epoch: 96 -> Loss: 1.01445615292 Epoch: 96 -> Test Accuracy: 53.2 [97, 60] loss: 1.113 [97, 120] loss: 1.092 [97, 180] loss: 1.117 [97, 240] loss: 1.101 [97, 300] loss: 1.110 [97, 360] loss: 1.111 Epoch: 97 -> Loss: 1.16818356514 Epoch: 97 -> Test Accuracy: 53.1 [98, 60] loss: 1.113 [98, 120] loss: 1.097 [98, 180] loss: 1.111 [98, 240] loss: 1.101 [98, 300] loss: 1.075 [98, 360] loss: 1.111 Epoch: 98 -> Loss: 1.11420583725 Epoch: 98 -> Test Accuracy: 53.73 [99, 60] loss: 1.104 [99, 120] loss: 1.107 [99, 180] loss: 1.111 [99, 240] loss: 1.095 [99, 300] loss: 1.083 [99, 360] loss: 1.095 Epoch: 99 -> Loss: 1.22115063667 Epoch: 99 -> Test Accuracy: 53.44 [100, 60] loss: 1.102 [100, 120] loss: 1.093 [100, 180] loss: 1.099 [100, 240] loss: 1.108 [100, 300] loss: 1.098 [100, 360] loss: 1.099 Epoch: 100 -> Loss: 1.22716128826 Epoch: 100 -> Test Accuracy: 53.4 Finished Training
In [10]:
# save variables
fm.save_variable([rot_block4_loss_log, rot_block4_test_accuracy_log,
block4_loss_log, block4_test_accuracy_log,
conv_block4_loss_log, conv_block4_test_accuracy_log], "4_block_net")
In [11]:
# rename files
fm.add_block_to_name(4, [100, 200])
5 Block RotNet¶
In [6]:
# initialize network
net_block5 = RN.RotNet(num_classes=4, num_conv_block=5, add_avg_pool=False)
In [7]:
# train network
rot_block5_loss_log, _, rot_block5_test_accuracy_log, _, _ = tr.adaptive_learning([0.1, 0.02, 0.004, 0.0008],
[60, 120, 160, 200], 0.9, 5e-4, net_block5, criterion, trainloader, None, testloader, rot=['90', '180', '270'])
functionalities/rotation.py:16: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. flip_idx = torch.range(trans_im.size(2) - 1, 0, -1).long() functionalities/rotation.py:31: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. vert_idx = torch.range(image.size(2) - 1, 0, -1).long() functionalities/rotation.py:35: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. hor_idx = torch.range(vert_im.size(1) - 1, 0, -1).long() functionalities/rotation.py:50: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end]. vert_idx = torch.range(image.size(2) - 1, 0, -1).long()
[1, 60] loss: 1.242 [1, 120] loss: 1.044 [1, 180] loss: 1.001 [1, 240] loss: 0.947 [1, 300] loss: 0.920 [1, 360] loss: 0.896 Epoch: 1 -> Loss: 0.777884483337 Epoch: 1 -> Test Accuracy: 65.8775 [2, 60] loss: 0.833 [2, 120] loss: 0.809 [2, 180] loss: 0.781 [2, 240] loss: 0.774 [2, 300] loss: 0.744 [2, 360] loss: 0.715 Epoch: 2 -> Loss: 0.648448169231 Epoch: 2 -> Test Accuracy: 73.33 [3, 60] loss: 0.674 [3, 120] loss: 0.659 [3, 180] loss: 0.653 [3, 240] loss: 0.636 [3, 300] loss: 0.637 [3, 360] loss: 0.629 Epoch: 3 -> Loss: 0.59783154726 Epoch: 3 -> Test Accuracy: 76.0625 [4, 60] loss: 0.609 [4, 120] loss: 0.590 [4, 180] loss: 0.575 [4, 240] loss: 0.564 [4, 300] loss: 0.570 [4, 360] loss: 0.568 Epoch: 4 -> Loss: 0.559232831001 Epoch: 4 -> Test Accuracy: 77.84 [5, 60] loss: 0.543 [5, 120] loss: 0.549 [5, 180] loss: 0.526 [5, 240] loss: 0.534 [5, 300] loss: 0.519 [5, 360] loss: 0.541 Epoch: 5 -> Loss: 0.41231456399 Epoch: 5 -> Test Accuracy: 80.7425 [6, 60] loss: 0.506 [6, 120] loss: 0.519 [6, 180] loss: 0.504 [6, 240] loss: 0.478 [6, 300] loss: 0.500 [6, 360] loss: 0.474 Epoch: 6 -> Loss: 0.504078507423 Epoch: 6 -> Test Accuracy: 79.7575 [7, 60] loss: 0.481 [7, 120] loss: 0.468 [7, 180] loss: 0.466 [7, 240] loss: 0.477 [7, 300] loss: 0.475 [7, 360] loss: 0.470 Epoch: 7 -> Loss: 0.382222265005 Epoch: 7 -> Test Accuracy: 80.2325 [8, 60] loss: 0.459 [8, 120] loss: 0.459 [8, 180] loss: 0.446 [8, 240] loss: 0.452 [8, 300] loss: 0.469 [8, 360] loss: 0.443 Epoch: 8 -> Loss: 0.461792856455 Epoch: 8 -> Test Accuracy: 81.92 [9, 60] loss: 0.429 [9, 120] loss: 0.447 [9, 180] loss: 0.456 [9, 240] loss: 0.428 [9, 300] loss: 0.447 [9, 360] loss: 0.441 Epoch: 9 -> Loss: 0.358080148697 Epoch: 9 -> Test Accuracy: 82.645 [10, 60] loss: 0.410 [10, 120] loss: 0.441 [10, 180] loss: 0.415 [10, 240] loss: 0.416 [10, 300] loss: 0.419 [10, 360] loss: 0.440 Epoch: 10 -> Loss: 0.329341083765 Epoch: 10 -> Test Accuracy: 83.3175 [11, 60] loss: 0.420 [11, 120] loss: 0.394 [11, 180] loss: 0.418 [11, 240] loss: 0.409 [11, 300] loss: 0.419 [11, 360] loss: 0.417 Epoch: 11 -> Loss: 0.457412064075 Epoch: 11 -> Test Accuracy: 82.8225 [12, 60] loss: 0.407 [12, 120] loss: 0.395 [12, 180] loss: 0.389 [12, 240] loss: 0.406 [12, 300] loss: 0.412 [12, 360] loss: 0.406 Epoch: 12 -> Loss: 0.360869258642 Epoch: 12 -> Test Accuracy: 83.8975 [13, 60] loss: 0.400 [13, 120] loss: 0.408 [13, 180] loss: 0.399 [13, 240] loss: 0.395 [13, 300] loss: 0.406 [13, 360] loss: 0.377 Epoch: 13 -> Loss: 0.46699398756 Epoch: 13 -> Test Accuracy: 84.5125 [14, 60] loss: 0.389 [14, 120] loss: 0.384 [14, 180] loss: 0.389 [14, 240] loss: 0.369 [14, 300] loss: 0.384 [14, 360] loss: 0.391 Epoch: 14 -> Loss: 0.378018081188 Epoch: 14 -> Test Accuracy: 84.855 [15, 60] loss: 0.387 [15, 120] loss: 0.388 [15, 180] loss: 0.378 [15, 240] loss: 0.380 [15, 300] loss: 0.377 [15, 360] loss: 0.373 Epoch: 15 -> Loss: 0.58175688982 Epoch: 15 -> Test Accuracy: 84.69 [16, 60] loss: 0.376 [16, 120] loss: 0.364 [16, 180] loss: 0.385 [16, 240] loss: 0.381 [16, 300] loss: 0.365 [16, 360] loss: 0.368 Epoch: 16 -> Loss: 0.40565481782 Epoch: 16 -> Test Accuracy: 83.62 [17, 60] loss: 0.350 [17, 120] loss: 0.374 [17, 180] loss: 0.367 [17, 240] loss: 0.372 [17, 300] loss: 0.370 [17, 360] loss: 0.375 Epoch: 17 -> Loss: 0.279655516148 Epoch: 17 -> Test Accuracy: 84.685 [18, 60] loss: 0.360 [18, 120] loss: 0.357 [18, 180] loss: 0.370 [18, 240] loss: 0.361 [18, 300] loss: 0.373 [18, 360] loss: 0.370 Epoch: 18 -> Loss: 0.338039547205 Epoch: 18 -> Test Accuracy: 85.03 [19, 60] loss: 0.349 [19, 120] loss: 0.356 [19, 180] loss: 0.353 [19, 240] loss: 0.354 [19, 300] loss: 0.368 [19, 360] loss: 0.366 Epoch: 19 -> Loss: 0.313662439585 Epoch: 19 -> Test Accuracy: 84.91 [20, 60] loss: 0.358 [20, 120] loss: 0.337 [20, 180] loss: 0.360 [20, 240] loss: 0.359 [20, 300] loss: 0.349 [20, 360] loss: 0.369 Epoch: 20 -> Loss: 0.408251821995 Epoch: 20 -> Test Accuracy: 84.8525 [21, 60] loss: 0.354 [21, 120] loss: 0.350 [21, 180] loss: 0.345 [21, 240] loss: 0.348 [21, 300] loss: 0.350 [21, 360] loss: 0.371 Epoch: 21 -> Loss: 0.233082175255 Epoch: 21 -> Test Accuracy: 85.165 [22, 60] loss: 0.333 [22, 120] loss: 0.354 [22, 180] loss: 0.347 [22, 240] loss: 0.363 [22, 300] loss: 0.339 [22, 360] loss: 0.367 Epoch: 22 -> Loss: 0.300312995911 Epoch: 22 -> Test Accuracy: 85.2375 [23, 60] loss: 0.340 [23, 120] loss: 0.348 [23, 180] loss: 0.341 [23, 240] loss: 0.342 [23, 300] loss: 0.349 [23, 360] loss: 0.343 Epoch: 23 -> Loss: 0.360540628433 Epoch: 23 -> Test Accuracy: 86.1225 [24, 60] loss: 0.337 [24, 120] loss: 0.346 [24, 180] loss: 0.337 [24, 240] loss: 0.350 [24, 300] loss: 0.339 [24, 360] loss: 0.349 Epoch: 24 -> Loss: 0.346710771322 Epoch: 24 -> Test Accuracy: 85.545 [25, 60] loss: 0.336 [25, 120] loss: 0.350 [25, 180] loss: 0.334 [25, 240] loss: 0.347 [25, 300] loss: 0.344 [25, 360] loss: 0.341 Epoch: 25 -> Loss: 0.445628076792 Epoch: 25 -> Test Accuracy: 85.6525 [26, 60] loss: 0.336 [26, 120] loss: 0.339 [26, 180] loss: 0.345 [26, 240] loss: 0.354 [26, 300] loss: 0.346 [26, 360] loss: 0.333 Epoch: 26 -> Loss: 0.357057720423 Epoch: 26 -> Test Accuracy: 85.71 [27, 60] loss: 0.332 [27, 120] loss: 0.351 [27, 180] loss: 0.335 [27, 240] loss: 0.339 [27, 300] loss: 0.328 [27, 360] loss: 0.330 Epoch: 27 -> Loss: 0.435141414404 Epoch: 27 -> Test Accuracy: 85.015 [28, 60] loss: 0.332 [28, 120] loss: 0.323 [28, 180] loss: 0.329 [28, 240] loss: 0.350 [28, 300] loss: 0.334 [28, 360] loss: 0.343 Epoch: 28 -> Loss: 0.30782777071 Epoch: 28 -> Test Accuracy: 85.1275 [29, 60] loss: 0.343 [29, 120] loss: 0.332 [29, 180] loss: 0.335 [29, 240] loss: 0.324 [29, 300] loss: 0.327 [29, 360] loss: 0.345 Epoch: 29 -> Loss: 0.337554395199 Epoch: 29 -> Test Accuracy: 86.305 [30, 60] loss: 0.309 [30, 120] loss: 0.323 [30, 180] loss: 0.355 [30, 240] loss: 0.330 [30, 300] loss: 0.338 [30, 360] loss: 0.334 Epoch: 30 -> Loss: 0.214543700218 Epoch: 30 -> Test Accuracy: 85.5475 [31, 60] loss: 0.330 [31, 120] loss: 0.334 [31, 180] loss: 0.318 [31, 240] loss: 0.338 [31, 300] loss: 0.331 [31, 360] loss: 0.341 Epoch: 31 -> Loss: 0.232660770416 Epoch: 31 -> Test Accuracy: 86.6125 [32, 60] loss: 0.319 [32, 120] loss: 0.329 [32, 180] loss: 0.330 [32, 240] loss: 0.336 [32, 300] loss: 0.328 [32, 360] loss: 0.335 Epoch: 32 -> Loss: 0.370219677687 Epoch: 32 -> Test Accuracy: 85.26 [33, 60] loss: 0.319 [33, 120] loss: 0.331 [33, 180] loss: 0.323 [33, 240] loss: 0.320 [33, 300] loss: 0.342 [33, 360] loss: 0.320 Epoch: 33 -> Loss: 0.436576515436 Epoch: 33 -> Test Accuracy: 86.49 [34, 60] loss: 0.305 [34, 120] loss: 0.324 [34, 180] loss: 0.335 [34, 240] loss: 0.329 [34, 300] loss: 0.325 [34, 360] loss: 0.323 Epoch: 34 -> Loss: 0.451147943735 Epoch: 34 -> Test Accuracy: 85.5075 [35, 60] loss: 0.318 [35, 120] loss: 0.307 [35, 180] loss: 0.336 [35, 240] loss: 0.323 [35, 300] loss: 0.327 [35, 360] loss: 0.322 Epoch: 35 -> Loss: 0.368377655745 Epoch: 35 -> Test Accuracy: 86.3525 [36, 60] loss: 0.317 [36, 120] loss: 0.308 [36, 180] loss: 0.318 [36, 240] loss: 0.328 [36, 300] loss: 0.335 [36, 360] loss: 0.316 Epoch: 36 -> Loss: 0.330079138279 Epoch: 36 -> Test Accuracy: 86.05 [37, 60] loss: 0.319 [37, 120] loss: 0.316 [37, 180] loss: 0.327 [37, 240] loss: 0.321 [37, 300] loss: 0.324 [37, 360] loss: 0.323 Epoch: 37 -> Loss: 0.301078379154 Epoch: 37 -> Test Accuracy: 86.45 [38, 60] loss: 0.302 [38, 120] loss: 0.310 [38, 180] loss: 0.333 [38, 240] loss: 0.319 [38, 300] loss: 0.326 [38, 360] loss: 0.340 Epoch: 38 -> Loss: 0.334967881441 Epoch: 38 -> Test Accuracy: 86.1275 [39, 60] loss: 0.312 [39, 120] loss: 0.329 [39, 180] loss: 0.315 [39, 240] loss: 0.326 [39, 300] loss: 0.309 [39, 360] loss: 0.324 Epoch: 39 -> Loss: 0.341781675816 Epoch: 39 -> Test Accuracy: 85.74 [40, 60] loss: 0.318 [40, 120] loss: 0.307 [40, 180] loss: 0.328 [40, 240] loss: 0.310 [40, 300] loss: 0.331 [40, 360] loss: 0.332 Epoch: 40 -> Loss: 0.517393827438 Epoch: 40 -> Test Accuracy: 85.225 [41, 60] loss: 0.308 [41, 120] loss: 0.326 [41, 180] loss: 0.333 [41, 240] loss: 0.306 [41, 300] loss: 0.324 [41, 360] loss: 0.315 Epoch: 41 -> Loss: 0.29975682497 Epoch: 41 -> Test Accuracy: 85.2725 [42, 60] loss: 0.309 [42, 120] loss: 0.311 [42, 180] loss: 0.318 [42, 240] loss: 0.312 [42, 300] loss: 0.321 [42, 360] loss: 0.331 Epoch: 42 -> Loss: 0.295903921127 Epoch: 42 -> Test Accuracy: 86.9525 [43, 60] loss: 0.304 [43, 120] loss: 0.327 [43, 180] loss: 0.303 [43, 240] loss: 0.327 [43, 300] loss: 0.322 [43, 360] loss: 0.330 Epoch: 43 -> Loss: 0.400561511517 Epoch: 43 -> Test Accuracy: 86.195 [44, 60] loss: 0.297 [44, 120] loss: 0.318 [44, 180] loss: 0.315 [44, 240] loss: 0.322 [44, 300] loss: 0.316 [44, 360] loss: 0.325 Epoch: 44 -> Loss: 0.30678999424 Epoch: 44 -> Test Accuracy: 85.77 [45, 60] loss: 0.303 [45, 120] loss: 0.317 [45, 180] loss: 0.321 [45, 240] loss: 0.314 [45, 300] loss: 0.306 [45, 360] loss: 0.328 Epoch: 45 -> Loss: 0.355306535959 Epoch: 45 -> Test Accuracy: 86.19 [46, 60] loss: 0.302 [46, 120] loss: 0.307 [46, 180] loss: 0.314 [46, 240] loss: 0.311 [46, 300] loss: 0.320 [46, 360] loss: 0.312 Epoch: 46 -> Loss: 0.31866440177 Epoch: 46 -> Test Accuracy: 85.425 [47, 60] loss: 0.312 [47, 120] loss: 0.320 [47, 180] loss: 0.314 [47, 240] loss: 0.302 [47, 300] loss: 0.320 [47, 360] loss: 0.314 Epoch: 47 -> Loss: 0.410285294056 Epoch: 47 -> Test Accuracy: 84.9575 [48, 60] loss: 0.304 [48, 120] loss: 0.310 [48, 180] loss: 0.321 [48, 240] loss: 0.313 [48, 300] loss: 0.316 [48, 360] loss: 0.317 Epoch: 48 -> Loss: 0.235875204206 Epoch: 48 -> Test Accuracy: 86.585 [49, 60] loss: 0.306 [49, 120] loss: 0.306 [49, 180] loss: 0.308 [49, 240] loss: 0.315 [49, 300] loss: 0.315 [49, 360] loss: 0.319 Epoch: 49 -> Loss: 0.37031275034 Epoch: 49 -> Test Accuracy: 86.4925 [50, 60] loss: 0.308 [50, 120] loss: 0.301 [50, 180] loss: 0.312 [50, 240] loss: 0.303 [50, 300] loss: 0.324 [50, 360] loss: 0.314 Epoch: 50 -> Loss: 0.308477640152 Epoch: 50 -> Test Accuracy: 85.06 [51, 60] loss: 0.299 [51, 120] loss: 0.304 [51, 180] loss: 0.319 [51, 240] loss: 0.318 [51, 300] loss: 0.311 [51, 360] loss: 0.305 Epoch: 51 -> Loss: 0.378709405661 Epoch: 51 -> Test Accuracy: 86.1075 [52, 60] loss: 0.306 [52, 120] loss: 0.311 [52, 180] loss: 0.311 [52, 240] loss: 0.314 [52, 300] loss: 0.306 [52, 360] loss: 0.325 Epoch: 52 -> Loss: 0.296770453453 Epoch: 52 -> Test Accuracy: 86.52 [53, 60] loss: 0.301 [53, 120] loss: 0.305 [53, 180] loss: 0.312 [53, 240] loss: 0.300 [53, 300] loss: 0.309 [53, 360] loss: 0.306 Epoch: 53 -> Loss: 0.296748191118 Epoch: 53 -> Test Accuracy: 86.5125 [54, 60] loss: 0.295 [54, 120] loss: 0.303 [54, 180] loss: 0.310 [54, 240] loss: 0.313 [54, 300] loss: 0.312 [54, 360] loss: 0.302 Epoch: 54 -> Loss: 0.283517181873 Epoch: 54 -> Test Accuracy: 86.0925 [55, 60] loss: 0.294 [55, 120] loss: 0.298 [55, 180] loss: 0.313 [55, 240] loss: 0.306 [55, 300] loss: 0.321 [55, 360] loss: 0.313 Epoch: 55 -> Loss: 0.341846287251 Epoch: 55 -> Test Accuracy: 86.625 [56, 60] loss: 0.283 [56, 120] loss: 0.304 [56, 180] loss: 0.305 [56, 240] loss: 0.316 [56, 300] loss: 0.315 [56, 360] loss: 0.319 Epoch: 56 -> Loss: 0.268593251705 Epoch: 56 -> Test Accuracy: 85.5575 [57, 60] loss: 0.296 [57, 120] loss: 0.292 [57, 180] loss: 0.307 [57, 240] loss: 0.309 [57, 300] loss: 0.323 [57, 360] loss: 0.306 Epoch: 57 -> Loss: 0.299621284008 Epoch: 57 -> Test Accuracy: 85.515 [58, 60] loss: 0.299 [58, 120] loss: 0.303 [58, 180] loss: 0.304 [58, 240] loss: 0.307 [58, 300] loss: 0.310 [58, 360] loss: 0.303 Epoch: 58 -> Loss: 0.391120016575 Epoch: 58 -> Test Accuracy: 87.0725 [59, 60] loss: 0.292 [59, 120] loss: 0.300 [59, 180] loss: 0.294 [59, 240] loss: 0.317 [59, 300] loss: 0.311 [59, 360] loss: 0.316 Epoch: 59 -> Loss: 0.31570148468 Epoch: 59 -> Test Accuracy: 87.095 [60, 60] loss: 0.307 [60, 120] loss: 0.296 [60, 180] loss: 0.302 [60, 240] loss: 0.302 [60, 300] loss: 0.296 [60, 360] loss: 0.324 Epoch: 60 -> Loss: 0.374734848738 Epoch: 60 -> Test Accuracy: 86.595 [61, 60] loss: 0.232 [61, 120] loss: 0.200 [61, 180] loss: 0.185 [61, 240] loss: 0.192 [61, 300] loss: 0.186 [61, 360] loss: 0.183 Epoch: 61 -> Loss: 0.157360211015 Epoch: 61 -> Test Accuracy: 90.7825 [62, 60] loss: 0.166 [62, 120] loss: 0.165 [62, 180] loss: 0.174 [62, 240] loss: 0.163 [62, 300] loss: 0.169 [62, 360] loss: 0.166 Epoch: 62 -> Loss: 0.198719024658 Epoch: 62 -> Test Accuracy: 91.1625 [63, 60] loss: 0.150 [63, 120] loss: 0.150 [63, 180] loss: 0.165 [63, 240] loss: 0.157 [63, 300] loss: 0.167 [63, 360] loss: 0.160 Epoch: 63 -> Loss: 0.135304674506 Epoch: 63 -> Test Accuracy: 90.925 [64, 60] loss: 0.139 [64, 120] loss: 0.154 [64, 180] loss: 0.154 [64, 240] loss: 0.157 [64, 300] loss: 0.150 [64, 360] loss: 0.159 Epoch: 64 -> Loss: 0.101378165185 Epoch: 64 -> Test Accuracy: 90.9575 [65, 60] loss: 0.146 [65, 120] loss: 0.155 [65, 180] loss: 0.152 [65, 240] loss: 0.152 [65, 300] loss: 0.146 [65, 360] loss: 0.151 Epoch: 65 -> Loss: 0.172543406487 Epoch: 65 -> Test Accuracy: 90.99 [66, 60] loss: 0.144 [66, 120] loss: 0.148 [66, 180] loss: 0.162 [66, 240] loss: 0.144 [66, 300] loss: 0.153 [66, 360] loss: 0.147 Epoch: 66 -> Loss: 0.122496888041 Epoch: 66 -> Test Accuracy: 91.1525 [67, 60] loss: 0.137 [67, 120] loss: 0.139 [67, 180] loss: 0.144 [67, 240] loss: 0.160 [67, 300] loss: 0.151 [67, 360] loss: 0.158 Epoch: 67 -> Loss: 0.145913600922 Epoch: 67 -> Test Accuracy: 91.03 [68, 60] loss: 0.137 [68, 120] loss: 0.142 [68, 180] loss: 0.148 [68, 240] loss: 0.160 [68, 300] loss: 0.144 [68, 360] loss: 0.151 Epoch: 68 -> Loss: 0.212457686663 Epoch: 68 -> Test Accuracy: 90.7575 [69, 60] loss: 0.142 [69, 120] loss: 0.144 [69, 180] loss: 0.144 [69, 240] loss: 0.152 [69, 300] loss: 0.149 [69, 360] loss: 0.151 Epoch: 69 -> Loss: 0.220518514514 Epoch: 69 -> Test Accuracy: 90.4725 [70, 60] loss: 0.145 [70, 120] loss: 0.156 [70, 180] loss: 0.145 [70, 240] loss: 0.144 [70, 300] loss: 0.145 [70, 360] loss: 0.151 Epoch: 70 -> Loss: 0.16164290905 Epoch: 70 -> Test Accuracy: 90.055 [71, 60] loss: 0.142 [71, 120] loss: 0.140 [71, 180] loss: 0.157 [71, 240] loss: 0.146 [71, 300] loss: 0.151 [71, 360] loss: 0.149 Epoch: 71 -> Loss: 0.0878470093012 Epoch: 71 -> Test Accuracy: 90.5975 [72, 60] loss: 0.136 [72, 120] loss: 0.141 [72, 180] loss: 0.149 [72, 240] loss: 0.148 [72, 300] loss: 0.150 [72, 360] loss: 0.158 Epoch: 72 -> Loss: 0.19203093648 Epoch: 72 -> Test Accuracy: 90.605 [73, 60] loss: 0.133 [73, 120] loss: 0.141 [73, 180] loss: 0.149 [73, 240] loss: 0.151 [73, 300] loss: 0.154 [73, 360] loss: 0.152 Epoch: 73 -> Loss: 0.157403796911 Epoch: 73 -> Test Accuracy: 90.5425 [74, 60] loss: 0.147 [74, 120] loss: 0.143 [74, 180] loss: 0.151 [74, 240] loss: 0.149 [74, 300] loss: 0.161 [74, 360] loss: 0.148 Epoch: 74 -> Loss: 0.0985128059983 Epoch: 74 -> Test Accuracy: 90.23 [75, 60] loss: 0.137 [75, 120] loss: 0.143 [75, 180] loss: 0.146 [75, 240] loss: 0.148 [75, 300] loss: 0.161 [75, 360] loss: 0.153 Epoch: 75 -> Loss: 0.192423030734 Epoch: 75 -> Test Accuracy: 90.2825 [76, 60] loss: 0.140 [76, 120] loss: 0.147 [76, 180] loss: 0.141 [76, 240] loss: 0.154 [76, 300] loss: 0.152 [76, 360] loss: 0.154 Epoch: 76 -> Loss: 0.21979098022 Epoch: 76 -> Test Accuracy: 90.7475 [77, 60] loss: 0.142 [77, 120] loss: 0.136 [77, 180] loss: 0.159 [77, 240] loss: 0.158 [77, 300] loss: 0.155 [77, 360] loss: 0.148 Epoch: 77 -> Loss: 0.198052495718 Epoch: 77 -> Test Accuracy: 89.8725 [78, 60] loss: 0.143 [78, 120] loss: 0.141 [78, 180] loss: 0.160 [78, 240] loss: 0.143 [78, 300] loss: 0.155 [78, 360] loss: 0.156 Epoch: 78 -> Loss: 0.118901535869 Epoch: 78 -> Test Accuracy: 90.0175 [79, 60] loss: 0.142 [79, 120] loss: 0.145 [79, 180] loss: 0.151 [79, 240] loss: 0.150 [79, 300] loss: 0.145 [79, 360] loss: 0.155 Epoch: 79 -> Loss: 0.174518898129 Epoch: 79 -> Test Accuracy: 90.3925 [80, 60] loss: 0.131 [80, 120] loss: 0.154 [80, 180] loss: 0.153 [80, 240] loss: 0.143 [80, 300] loss: 0.156 [80, 360] loss: 0.151 Epoch: 80 -> Loss: 0.176137581468 Epoch: 80 -> Test Accuracy: 90.425 [81, 60] loss: 0.139 [81, 120] loss: 0.153 [81, 180] loss: 0.142 [81, 240] loss: 0.143 [81, 300] loss: 0.150 [81, 360] loss: 0.158 Epoch: 81 -> Loss: 0.115477837622 Epoch: 81 -> Test Accuracy: 90.4975 [82, 60] loss: 0.137 [82, 120] loss: 0.148 [82, 180] loss: 0.149 [82, 240] loss: 0.155 [82, 300] loss: 0.151 [82, 360] loss: 0.160 Epoch: 82 -> Loss: 0.0695660114288 Epoch: 82 -> Test Accuracy: 90.59 [83, 60] loss: 0.133 [83, 120] loss: 0.147 [83, 180] loss: 0.137 [83, 240] loss: 0.152 [83, 300] loss: 0.156 [83, 360] loss: 0.160 Epoch: 83 -> Loss: 0.213991358876 Epoch: 83 -> Test Accuracy: 90.3625 [84, 60] loss: 0.133 [84, 120] loss: 0.142 [84, 180] loss: 0.141 [84, 240] loss: 0.151 [84, 300] loss: 0.157 [84, 360] loss: 0.156 Epoch: 84 -> Loss: 0.158650770783 Epoch: 84 -> Test Accuracy: 89.905 [85, 60] loss: 0.144 [85, 120] loss: 0.142 [85, 180] loss: 0.154 [85, 240] loss: 0.142 [85, 300] loss: 0.153 [85, 360] loss: 0.155 Epoch: 85 -> Loss: 0.133175000548 Epoch: 85 -> Test Accuracy: 89.7075 [86, 60] loss: 0.137 [86, 120] loss: 0.145 [86, 180] loss: 0.148 [86, 240] loss: 0.143 [86, 300] loss: 0.152 [86, 360] loss: 0.154 Epoch: 86 -> Loss: 0.0884124040604 Epoch: 86 -> Test Accuracy: 90.0125 [87, 60] loss: 0.142 [87, 120] loss: 0.138 [87, 180] loss: 0.149 [87, 240] loss: 0.154 [87, 300] loss: 0.144 [87, 360] loss: 0.151 Epoch: 87 -> Loss: 0.109030939639 Epoch: 87 -> Test Accuracy: 89.9325 [88, 60] loss: 0.138 [88, 120] loss: 0.146 [88, 180] loss: 0.143 [88, 240] loss: 0.150 [88, 300] loss: 0.146 [88, 360] loss: 0.149 Epoch: 88 -> Loss: 0.149111643434 Epoch: 88 -> Test Accuracy: 90.465 [89, 60] loss: 0.135 [89, 120] loss: 0.141 [89, 180] loss: 0.142 [89, 240] loss: 0.149 [89, 300] loss: 0.145 [89, 360] loss: 0.151 Epoch: 89 -> Loss: 0.146894484758 Epoch: 89 -> Test Accuracy: 89.8125 [90, 60] loss: 0.139 [90, 120] loss: 0.140 [90, 180] loss: 0.145 [90, 240] loss: 0.148 [90, 300] loss: 0.152 [90, 360] loss: 0.137 Epoch: 90 -> Loss: 0.133741512895 Epoch: 90 -> Test Accuracy: 89.9725 [91, 60] loss: 0.141 [91, 120] loss: 0.137 [91, 180] loss: 0.144 [91, 240] loss: 0.138 [91, 300] loss: 0.153 [91, 360] loss: 0.149 Epoch: 91 -> Loss: 0.134075343609 Epoch: 91 -> Test Accuracy: 90.3725 [92, 60] loss: 0.129 [92, 120] loss: 0.139 [92, 180] loss: 0.141 [92, 240] loss: 0.136 [92, 300] loss: 0.148 [92, 360] loss: 0.147 Epoch: 92 -> Loss: 0.345908850431 Epoch: 92 -> Test Accuracy: 90.355 [93, 60] loss: 0.125 [93, 120] loss: 0.130 [93, 180] loss: 0.154 [93, 240] loss: 0.137 [93, 300] loss: 0.155 [93, 360] loss: 0.148 Epoch: 93 -> Loss: 0.207707047462 Epoch: 93 -> Test Accuracy: 90.1275 [94, 60] loss: 0.141 [94, 120] loss: 0.142 [94, 180] loss: 0.140 [94, 240] loss: 0.134 [94, 300] loss: 0.147 [94, 360] loss: 0.152 Epoch: 94 -> Loss: 0.155148491263 Epoch: 94 -> Test Accuracy: 90.35 [95, 60] loss: 0.133 [95, 120] loss: 0.136 [95, 180] loss: 0.142 [95, 240] loss: 0.152 [95, 300] loss: 0.148 [95, 360] loss: 0.147 Epoch: 95 -> Loss: 0.128780096769 Epoch: 95 -> Test Accuracy: 90.355 [96, 60] loss: 0.127 [96, 120] loss: 0.137 [96, 180] loss: 0.142 [96, 240] loss: 0.136 [96, 300] loss: 0.152 [96, 360] loss: 0.144 Epoch: 96 -> Loss: 0.197761058807 Epoch: 96 -> Test Accuracy: 90.1825 [97, 60] loss: 0.134 [97, 120] loss: 0.130 [97, 180] loss: 0.139 [97, 240] loss: 0.137 [97, 300] loss: 0.151 [97, 360] loss: 0.146 Epoch: 97 -> Loss: 0.144468128681 Epoch: 97 -> Test Accuracy: 89.9425 [98, 60] loss: 0.127 [98, 120] loss: 0.130 [98, 180] loss: 0.137 [98, 240] loss: 0.145 [98, 300] loss: 0.146 [98, 360] loss: 0.142 Epoch: 98 -> Loss: 0.123912729323 Epoch: 98 -> Test Accuracy: 90.1125 [99, 60] loss: 0.136 [99, 120] loss: 0.136 [99, 180] loss: 0.143 [99, 240] loss: 0.136 [99, 300] loss: 0.146 [99, 360] loss: 0.144 Epoch: 99 -> Loss: 0.244579985738 Epoch: 99 -> Test Accuracy: 90.3475 [100, 60] loss: 0.131 [100, 120] loss: 0.144 [100, 180] loss: 0.138 [100, 240] loss: 0.143 [100, 300] loss: 0.143 [100, 360] loss: 0.147 Epoch: 100 -> Loss: 0.0971162691712 Epoch: 100 -> Test Accuracy: 90.275 [101, 60] loss: 0.126 [101, 120] loss: 0.135 [101, 180] loss: 0.141 [101, 240] loss: 0.142 [101, 300] loss: 0.141 [101, 360] loss: 0.152 Epoch: 101 -> Loss: 0.122770212591 Epoch: 101 -> Test Accuracy: 90.2575 [102, 60] loss: 0.132 [102, 120] loss: 0.133 [102, 180] loss: 0.135 [102, 240] loss: 0.143 [102, 300] loss: 0.138 [102, 360] loss: 0.141 Epoch: 102 -> Loss: 0.149077519774 Epoch: 102 -> Test Accuracy: 90.205 [103, 60] loss: 0.134 [103, 120] loss: 0.136 [103, 180] loss: 0.133 [103, 240] loss: 0.139 [103, 300] loss: 0.144 [103, 360] loss: 0.151 Epoch: 103 -> Loss: 0.156240969896 Epoch: 103 -> Test Accuracy: 90.26 [104, 60] loss: 0.131 [104, 120] loss: 0.137 [104, 180] loss: 0.142 [104, 240] loss: 0.138 [104, 300] loss: 0.147 [104, 360] loss: 0.146 Epoch: 104 -> Loss: 0.235925555229 Epoch: 104 -> Test Accuracy: 90.0275 [105, 60] loss: 0.124 [105, 120] loss: 0.141 [105, 180] loss: 0.131 [105, 240] loss: 0.143 [105, 300] loss: 0.140 [105, 360] loss: 0.144 Epoch: 105 -> Loss: 0.107051491737 Epoch: 105 -> Test Accuracy: 90.52 [106, 60] loss: 0.128 [106, 120] loss: 0.129 [106, 180] loss: 0.137 [106, 240] loss: 0.150 [106, 300] loss: 0.133 [106, 360] loss: 0.137 Epoch: 106 -> Loss: 0.121595367789 Epoch: 106 -> Test Accuracy: 90.45 [107, 60] loss: 0.128 [107, 120] loss: 0.135 [107, 180] loss: 0.132 [107, 240] loss: 0.135 [107, 300] loss: 0.132 [107, 360] loss: 0.147 Epoch: 107 -> Loss: 0.266189336777 Epoch: 107 -> Test Accuracy: 90.4675 [108, 60] loss: 0.133 [108, 120] loss: 0.125 [108, 180] loss: 0.138 [108, 240] loss: 0.141 [108, 300] loss: 0.146 [108, 360] loss: 0.141 Epoch: 108 -> Loss: 0.0670185759664 Epoch: 108 -> Test Accuracy: 90.3825 [109, 60] loss: 0.122 [109, 120] loss: 0.130 [109, 180] loss: 0.135 [109, 240] loss: 0.138 [109, 300] loss: 0.133 [109, 360] loss: 0.151 Epoch: 109 -> Loss: 0.18507103622 Epoch: 109 -> Test Accuracy: 89.91 [110, 60] loss: 0.117 [110, 120] loss: 0.137 [110, 180] loss: 0.137 [110, 240] loss: 0.142 [110, 300] loss: 0.139 [110, 360] loss: 0.133 Epoch: 110 -> Loss: 0.130577296019 Epoch: 110 -> Test Accuracy: 90.96 [111, 60] loss: 0.127 [111, 120] loss: 0.134 [111, 180] loss: 0.133 [111, 240] loss: 0.131 [111, 300] loss: 0.133 [111, 360] loss: 0.147 Epoch: 111 -> Loss: 0.106502607465 Epoch: 111 -> Test Accuracy: 90.0725 [112, 60] loss: 0.120 [112, 120] loss: 0.131 [112, 180] loss: 0.142 [112, 240] loss: 0.130 [112, 300] loss: 0.144 [112, 360] loss: 0.136 Epoch: 112 -> Loss: 0.0815871208906 Epoch: 112 -> Test Accuracy: 90.205 [113, 60] loss: 0.126 [113, 120] loss: 0.138 [113, 180] loss: 0.131 [113, 240] loss: 0.131 [113, 300] loss: 0.142 [113, 360] loss: 0.142 Epoch: 113 -> Loss: 0.189298853278 Epoch: 113 -> Test Accuracy: 90.42 [114, 60] loss: 0.124 [114, 120] loss: 0.125 [114, 180] loss: 0.133 [114, 240] loss: 0.139 [114, 300] loss: 0.143 [114, 360] loss: 0.140 Epoch: 114 -> Loss: 0.246730536222 Epoch: 114 -> Test Accuracy: 90.03 [115, 60] loss: 0.124 [115, 120] loss: 0.122 [115, 180] loss: 0.131 [115, 240] loss: 0.145 [115, 300] loss: 0.134 [115, 360] loss: 0.147 Epoch: 115 -> Loss: 0.148653298616 Epoch: 115 -> Test Accuracy: 90.3875 [116, 60] loss: 0.123 [116, 120] loss: 0.126 [116, 180] loss: 0.135 [116, 240] loss: 0.129 [116, 300] loss: 0.142 [116, 360] loss: 0.143 Epoch: 116 -> Loss: 0.190787643194 Epoch: 116 -> Test Accuracy: 89.7875 [117, 60] loss: 0.121 [117, 120] loss: 0.130 [117, 180] loss: 0.136 [117, 240] loss: 0.142 [117, 300] loss: 0.132 [117, 360] loss: 0.136 Epoch: 117 -> Loss: 0.130016759038 Epoch: 117 -> Test Accuracy: 90.53 [118, 60] loss: 0.124 [118, 120] loss: 0.126 [118, 180] loss: 0.139 [118, 240] loss: 0.135 [118, 300] loss: 0.138 [118, 360] loss: 0.133 Epoch: 118 -> Loss: 0.168916806579 Epoch: 118 -> Test Accuracy: 90.1275 [119, 60] loss: 0.123 [119, 120] loss: 0.133 [119, 180] loss: 0.132 [119, 240] loss: 0.135 [119, 300] loss: 0.131 [119, 360] loss: 0.128 Epoch: 119 -> Loss: 0.16822052002 Epoch: 119 -> Test Accuracy: 90.69 [120, 60] loss: 0.125 [120, 120] loss: 0.129 [120, 180] loss: 0.121 [120, 240] loss: 0.139 [120, 300] loss: 0.136 [120, 360] loss: 0.137 Epoch: 120 -> Loss: 0.0735254511237 Epoch: 120 -> Test Accuracy: 90.34 [121, 60] loss: 0.102 [121, 120] loss: 0.072 [121, 180] loss: 0.068 [121, 240] loss: 0.065 [121, 300] loss: 0.065 [121, 360] loss: 0.062 Epoch: 121 -> Loss: 0.0345267057419 Epoch: 121 -> Test Accuracy: 92.1575 [122, 60] loss: 0.056 [122, 120] loss: 0.051 [122, 180] loss: 0.053 [122, 240] loss: 0.055 [122, 300] loss: 0.055 [122, 360] loss: 0.058 Epoch: 122 -> Loss: 0.0269163753837 Epoch: 122 -> Test Accuracy: 92.3825 [123, 60] loss: 0.048 [123, 120] loss: 0.050 [123, 180] loss: 0.049 [123, 240] loss: 0.049 [123, 300] loss: 0.050 [123, 360] loss: 0.049 Epoch: 123 -> Loss: 0.0529579408467 Epoch: 123 -> Test Accuracy: 92.18 [124, 60] loss: 0.041 [124, 120] loss: 0.038 [124, 180] loss: 0.043 [124, 240] loss: 0.047 [124, 300] loss: 0.043 [124, 360] loss: 0.043 Epoch: 124 -> Loss: 0.0119058378041 Epoch: 124 -> Test Accuracy: 92.0225 [125, 60] loss: 0.043 [125, 120] loss: 0.038 [125, 180] loss: 0.041 [125, 240] loss: 0.043 [125, 300] loss: 0.039 [125, 360] loss: 0.038 Epoch: 125 -> Loss: 0.084841221571 Epoch: 125 -> Test Accuracy: 92.2125 [126, 60] loss: 0.034 [126, 120] loss: 0.033 [126, 180] loss: 0.038 [126, 240] loss: 0.039 [126, 300] loss: 0.041 [126, 360] loss: 0.040 Epoch: 126 -> Loss: 0.0972911864519 Epoch: 126 -> Test Accuracy: 92.1425 [127, 60] loss: 0.036 [127, 120] loss: 0.034 [127, 180] loss: 0.034 [127, 240] loss: 0.035 [127, 300] loss: 0.037 [127, 360] loss: 0.035 Epoch: 127 -> Loss: 0.00685061747208 Epoch: 127 -> Test Accuracy: 92.0225 [128, 60] loss: 0.033 [128, 120] loss: 0.035 [128, 180] loss: 0.033 [128, 240] loss: 0.034 [128, 300] loss: 0.037 [128, 360] loss: 0.036 Epoch: 128 -> Loss: 0.0581314563751 Epoch: 128 -> Test Accuracy: 91.9325 [129, 60] loss: 0.033 [129, 120] loss: 0.032 [129, 180] loss: 0.034 [129, 240] loss: 0.034 [129, 300] loss: 0.033 [129, 360] loss: 0.033 Epoch: 129 -> Loss: 0.0307410992682 Epoch: 129 -> Test Accuracy: 92.1625 [130, 60] loss: 0.032 [130, 120] loss: 0.030 [130, 180] loss: 0.031 [130, 240] loss: 0.034 [130, 300] loss: 0.032 [130, 360] loss: 0.031 Epoch: 130 -> Loss: 0.0263896342367 Epoch: 130 -> Test Accuracy: 92.0975 [131, 60] loss: 0.032 [131, 120] loss: 0.032 [131, 180] loss: 0.029 [131, 240] loss: 0.032 [131, 300] loss: 0.035 [131, 360] loss: 0.032 Epoch: 131 -> Loss: 0.0555750355124 Epoch: 131 -> Test Accuracy: 92.0025 [132, 60] loss: 0.032 [132, 120] loss: 0.030 [132, 180] loss: 0.029 [132, 240] loss: 0.031 [132, 300] loss: 0.030 [132, 360] loss: 0.033 Epoch: 132 -> Loss: 0.0544689483941 Epoch: 132 -> Test Accuracy: 92.13 [133, 60] loss: 0.029 [133, 120] loss: 0.029 [133, 180] loss: 0.028 [133, 240] loss: 0.028 [133, 300] loss: 0.030 [133, 360] loss: 0.033 Epoch: 133 -> Loss: 0.0429929457605 Epoch: 133 -> Test Accuracy: 91.7125 [134, 60] loss: 0.031 [134, 120] loss: 0.031 [134, 180] loss: 0.028 [134, 240] loss: 0.028 [134, 300] loss: 0.026 [134, 360] loss: 0.032 Epoch: 134 -> Loss: 0.0229850467294 Epoch: 134 -> Test Accuracy: 92.115 [135, 60] loss: 0.029 [135, 120] loss: 0.028 [135, 180] loss: 0.030 [135, 240] loss: 0.027 [135, 300] loss: 0.027 [135, 360] loss: 0.031 Epoch: 135 -> Loss: 0.0190198067576 Epoch: 135 -> Test Accuracy: 91.7575 [136, 60] loss: 0.029 [136, 120] loss: 0.028 [136, 180] loss: 0.028 [136, 240] loss: 0.028 [136, 300] loss: 0.028 [136, 360] loss: 0.027 Epoch: 136 -> Loss: 0.0446188338101 Epoch: 136 -> Test Accuracy: 91.91 [137, 60] loss: 0.027 [137, 120] loss: 0.027 [137, 180] loss: 0.026 [137, 240] loss: 0.027 [137, 300] loss: 0.031 [137, 360] loss: 0.028 Epoch: 137 -> Loss: 0.0321652516723 Epoch: 137 -> Test Accuracy: 92.0175 [138, 60] loss: 0.025 [138, 120] loss: 0.027 [138, 180] loss: 0.028 [138, 240] loss: 0.027 [138, 300] loss: 0.028 [138, 360] loss: 0.029 Epoch: 138 -> Loss: 0.0293081011623 Epoch: 138 -> Test Accuracy: 91.9175 [139, 60] loss: 0.025 [139, 120] loss: 0.028 [139, 180] loss: 0.026 [139, 240] loss: 0.026 [139, 300] loss: 0.029 [139, 360] loss: 0.028 Epoch: 139 -> Loss: 0.0131450919434 Epoch: 139 -> Test Accuracy: 91.7675 [140, 60] loss: 0.026 [140, 120] loss: 0.025 [140, 180] loss: 0.026 [140, 240] loss: 0.030 [140, 300] loss: 0.028 [140, 360] loss: 0.028 Epoch: 140 -> Loss: 0.0177784115076 Epoch: 140 -> Test Accuracy: 91.5675 [141, 60] loss: 0.025 [141, 120] loss: 0.028 [141, 180] loss: 0.025 [141, 240] loss: 0.027 [141, 300] loss: 0.031 [141, 360] loss: 0.029 Epoch: 141 -> Loss: 0.021506762132 Epoch: 141 -> Test Accuracy: 91.7625 [142, 60] loss: 0.024 [142, 120] loss: 0.026 [142, 180] loss: 0.025 [142, 240] loss: 0.029 [142, 300] loss: 0.032 [142, 360] loss: 0.030 Epoch: 142 -> Loss: 0.03966383636 Epoch: 142 -> Test Accuracy: 91.785 [143, 60] loss: 0.026 [143, 120] loss: 0.030 [143, 180] loss: 0.028 [143, 240] loss: 0.028 [143, 300] loss: 0.028 [143, 360] loss: 0.026 Epoch: 143 -> Loss: 0.01487852633 Epoch: 143 -> Test Accuracy: 91.7 [144, 60] loss: 0.026 [144, 120] loss: 0.027 [144, 180] loss: 0.026 [144, 240] loss: 0.025 [144, 300] loss: 0.027 [144, 360] loss: 0.030 Epoch: 144 -> Loss: 0.0132517497987 Epoch: 144 -> Test Accuracy: 91.6375 [145, 60] loss: 0.028 [145, 120] loss: 0.027 [145, 180] loss: 0.025 [145, 240] loss: 0.028 [145, 300] loss: 0.029 [145, 360] loss: 0.030 Epoch: 145 -> Loss: 0.0381929054856 Epoch: 145 -> Test Accuracy: 91.5125 [146, 60] loss: 0.026 [146, 120] loss: 0.027 [146, 180] loss: 0.026 [146, 240] loss: 0.024 [146, 300] loss: 0.026 [146, 360] loss: 0.026 Epoch: 146 -> Loss: 0.0259801056236 Epoch: 146 -> Test Accuracy: 91.45 [147, 60] loss: 0.027 [147, 120] loss: 0.024 [147, 180] loss: 0.030 [147, 240] loss: 0.027 [147, 300] loss: 0.026 [147, 360] loss: 0.027 Epoch: 147 -> Loss: 0.0223569516093 Epoch: 147 -> Test Accuracy: 91.49 [148, 60] loss: 0.024 [148, 120] loss: 0.025 [148, 180] loss: 0.026 [148, 240] loss: 0.028 [148, 300] loss: 0.030 [148, 360] loss: 0.028 Epoch: 148 -> Loss: 0.0316341593862 Epoch: 148 -> Test Accuracy: 91.5475 [149, 60] loss: 0.028 [149, 120] loss: 0.027 [149, 180] loss: 0.027 [149, 240] loss: 0.027 [149, 300] loss: 0.027 [149, 360] loss: 0.030 Epoch: 149 -> Loss: 0.054685793817 Epoch: 149 -> Test Accuracy: 91.7275 [150, 60] loss: 0.027 [150, 120] loss: 0.027 [150, 180] loss: 0.027 [150, 240] loss: 0.027 [150, 300] loss: 0.029 [150, 360] loss: 0.030 Epoch: 150 -> Loss: 0.0487078540027 Epoch: 150 -> Test Accuracy: 91.6125 [151, 60] loss: 0.026 [151, 120] loss: 0.026 [151, 180] loss: 0.022 [151, 240] loss: 0.029 [151, 300] loss: 0.033 [151, 360] loss: 0.034 Epoch: 151 -> Loss: 0.0422250255942 Epoch: 151 -> Test Accuracy: 91.6625 [152, 60] loss: 0.025 [152, 120] loss: 0.024 [152, 180] loss: 0.029 [152, 240] loss: 0.028 [152, 300] loss: 0.030 [152, 360] loss: 0.030 Epoch: 152 -> Loss: 0.0139181539416 Epoch: 152 -> Test Accuracy: 91.6 [153, 60] loss: 0.026 [153, 120] loss: 0.024 [153, 180] loss: 0.026 [153, 240] loss: 0.028 [153, 300] loss: 0.026 [153, 360] loss: 0.027 Epoch: 153 -> Loss: 0.0170603133738 Epoch: 153 -> Test Accuracy: 91.595 [154, 60] loss: 0.025 [154, 120] loss: 0.030 [154, 180] loss: 0.027 [154, 240] loss: 0.029 [154, 300] loss: 0.028 [154, 360] loss: 0.030 Epoch: 154 -> Loss: 0.0775458365679 Epoch: 154 -> Test Accuracy: 91.6475 [155, 60] loss: 0.028 [155, 120] loss: 0.028 [155, 180] loss: 0.026 [155, 240] loss: 0.027 [155, 300] loss: 0.031 [155, 360] loss: 0.031 Epoch: 155 -> Loss: 0.0347133874893 Epoch: 155 -> Test Accuracy: 91.7375 [156, 60] loss: 0.030 [156, 120] loss: 0.026 [156, 180] loss: 0.029 [156, 240] loss: 0.029 [156, 300] loss: 0.030 [156, 360] loss: 0.027 Epoch: 156 -> Loss: 0.0320833846927 Epoch: 156 -> Test Accuracy: 91.77 [157, 60] loss: 0.022 [157, 120] loss: 0.027 [157, 180] loss: 0.027 [157, 240] loss: 0.028 [157, 300] loss: 0.031 [157, 360] loss: 0.030 Epoch: 157 -> Loss: 0.0669270306826 Epoch: 157 -> Test Accuracy: 91.5275 [158, 60] loss: 0.027 [158, 120] loss: 0.029 [158, 180] loss: 0.027 [158, 240] loss: 0.028 [158, 300] loss: 0.027 [158, 360] loss: 0.034 Epoch: 158 -> Loss: 0.0528500676155 Epoch: 158 -> Test Accuracy: 91.2725 [159, 60] loss: 0.027 [159, 120] loss: 0.029 [159, 180] loss: 0.027 [159, 240] loss: 0.031 [159, 300] loss: 0.028 [159, 360] loss: 0.033 Epoch: 159 -> Loss: 0.0233434811234 Epoch: 159 -> Test Accuracy: 91.66 [160, 60] loss: 0.026 [160, 120] loss: 0.027 [160, 180] loss: 0.028 [160, 240] loss: 0.029 [160, 300] loss: 0.028 [160, 360] loss: 0.030 Epoch: 160 -> Loss: 0.0234837271273 Epoch: 160 -> Test Accuracy: 91.5625 [161, 60] loss: 0.020 [161, 120] loss: 0.020 [161, 180] loss: 0.016 [161, 240] loss: 0.015 [161, 300] loss: 0.015 [161, 360] loss: 0.015 Epoch: 161 -> Loss: 0.0120748449117 Epoch: 161 -> Test Accuracy: 92.1325 [162, 60] loss: 0.013 [162, 120] loss: 0.013 [162, 180] loss: 0.012 [162, 240] loss: 0.010 [162, 300] loss: 0.012 [162, 360] loss: 0.012 Epoch: 162 -> Loss: 0.00365521688946 Epoch: 162 -> Test Accuracy: 92.1675 [163, 60] loss: 0.010 [163, 120] loss: 0.010 [163, 180] loss: 0.011 [163, 240] loss: 0.011 [163, 300] loss: 0.011 [163, 360] loss: 0.010 Epoch: 163 -> Loss: 0.00734525918961 Epoch: 163 -> Test Accuracy: 92.295 [164, 60] loss: 0.009 [164, 120] loss: 0.010 [164, 180] loss: 0.009 [164, 240] loss: 0.011 [164, 300] loss: 0.010 [164, 360] loss: 0.009 Epoch: 164 -> Loss: 0.0237358827144 Epoch: 164 -> Test Accuracy: 92.325 [165, 60] loss: 0.010 [165, 120] loss: 0.009 [165, 180] loss: 0.009 [165, 240] loss: 0.009 [165, 300] loss: 0.008 [165, 360] loss: 0.009 Epoch: 165 -> Loss: 0.0161010064185 Epoch: 165 -> Test Accuracy: 92.3025 [166, 60] loss: 0.008 [166, 120] loss: 0.008 [166, 180] loss: 0.008 [166, 240] loss: 0.009 [166, 300] loss: 0.008 [166, 360] loss: 0.009 Epoch: 166 -> Loss: 0.00349425966851 Epoch: 166 -> Test Accuracy: 92.325 [167, 60] loss: 0.007 [167, 120] loss: 0.008 [167, 180] loss: 0.007 [167, 240] loss: 0.008 [167, 300] loss: 0.008 [167, 360] loss: 0.008 Epoch: 167 -> Loss: 0.00289918854833 Epoch: 167 -> Test Accuracy: 92.4 [168, 60] loss: 0.007 [168, 120] loss: 0.007 [168, 180] loss: 0.008 [168, 240] loss: 0.007 [168, 300] loss: 0.007 [168, 360] loss: 0.007 Epoch: 168 -> Loss: 0.0039829374291 Epoch: 168 -> Test Accuracy: 92.2675 [169, 60] loss: 0.008 [169, 120] loss: 0.008 [169, 180] loss: 0.007 [169, 240] loss: 0.008 [169, 300] loss: 0.008 [169, 360] loss: 0.008 Epoch: 169 -> Loss: 0.00722488900647 Epoch: 169 -> Test Accuracy: 92.1725 [170, 60] loss: 0.007 [170, 120] loss: 0.008 [170, 180] loss: 0.008 [170, 240] loss: 0.007 [170, 300] loss: 0.008 [170, 360] loss: 0.008 Epoch: 170 -> Loss: 0.00111298705451 Epoch: 170 -> Test Accuracy: 92.29 [171, 60] loss: 0.007 [171, 120] loss: 0.006 [171, 180] loss: 0.007 [171, 240] loss: 0.007 [171, 300] loss: 0.007 [171, 360] loss: 0.006 Epoch: 171 -> Loss: 0.0110035929829 Epoch: 171 -> Test Accuracy: 92.3 [172, 60] loss: 0.007 [172, 120] loss: 0.006 [172, 180] loss: 0.006 [172, 240] loss: 0.007 [172, 300] loss: 0.007 [172, 360] loss: 0.007 Epoch: 172 -> Loss: 0.00579565856606 Epoch: 172 -> Test Accuracy: 92.31 [173, 60] loss: 0.007 [173, 120] loss: 0.006 [173, 180] loss: 0.007 [173, 240] loss: 0.008 [173, 300] loss: 0.005 [173, 360] loss: 0.007 Epoch: 173 -> Loss: 0.00440714554861 Epoch: 173 -> Test Accuracy: 92.3225 [174, 60] loss: 0.006 [174, 120] loss: 0.006 [174, 180] loss: 0.006 [174, 240] loss: 0.006 [174, 300] loss: 0.006 [174, 360] loss: 0.006 Epoch: 174 -> Loss: 0.00424929708242 Epoch: 174 -> Test Accuracy: 92.265 [175, 60] loss: 0.006 [175, 120] loss: 0.006 [175, 180] loss: 0.006 [175, 240] loss: 0.006 [175, 300] loss: 0.005 [175, 360] loss: 0.007 Epoch: 175 -> Loss: 0.0104875480756 Epoch: 175 -> Test Accuracy: 92.2475 [176, 60] loss: 0.006 [176, 120] loss: 0.006 [176, 180] loss: 0.006 [176, 240] loss: 0.006 [176, 300] loss: 0.006 [176, 360] loss: 0.006 Epoch: 176 -> Loss: 0.00488089257851 Epoch: 176 -> Test Accuracy: 92.185 [177, 60] loss: 0.006 [177, 120] loss: 0.006 [177, 180] loss: 0.006 [177, 240] loss: 0.006 [177, 300] loss: 0.005 [177, 360] loss: 0.006 Epoch: 177 -> Loss: 0.00252343830653 Epoch: 177 -> Test Accuracy: 92.2425 [178, 60] loss: 0.005 [178, 120] loss: 0.005 [178, 180] loss: 0.006 [178, 240] loss: 0.005 [178, 300] loss: 0.006 [178, 360] loss: 0.005 Epoch: 178 -> Loss: 0.00848292559385 Epoch: 178 -> Test Accuracy: 92.3325 [179, 60] loss: 0.005 [179, 120] loss: 0.006 [179, 180] loss: 0.006 [179, 240] loss: 0.006 [179, 300] loss: 0.005 [179, 360] loss: 0.006 Epoch: 179 -> Loss: 0.00370348757133 Epoch: 179 -> Test Accuracy: 92.22 [180, 60] loss: 0.005 [180, 120] loss: 0.005 [180, 180] loss: 0.006 [180, 240] loss: 0.006 [180, 300] loss: 0.006 [180, 360] loss: 0.006 Epoch: 180 -> Loss: 0.00691954931244 Epoch: 180 -> Test Accuracy: 92.2775 [181, 60] loss: 0.005 [181, 120] loss: 0.005 [181, 180] loss: 0.004 [181, 240] loss: 0.005 [181, 300] loss: 0.005 [181, 360] loss: 0.005 Epoch: 181 -> Loss: 0.00417579058558 Epoch: 181 -> Test Accuracy: 92.2975 [182, 60] loss: 0.006 [182, 120] loss: 0.006 [182, 180] loss: 0.005 [182, 240] loss: 0.005 [182, 300] loss: 0.005 [182, 360] loss: 0.005 Epoch: 182 -> Loss: 0.00415513291955 Epoch: 182 -> Test Accuracy: 92.23 [183, 60] loss: 0.005 [183, 120] loss: 0.005 [183, 180] loss: 0.005 [183, 240] loss: 0.005 [183, 300] loss: 0.005 [183, 360] loss: 0.005 Epoch: 183 -> Loss: 0.00320652802475 Epoch: 183 -> Test Accuracy: 92.2675 [184, 60] loss: 0.005 [184, 120] loss: 0.006 [184, 180] loss: 0.006 [184, 240] loss: 0.005 [184, 300] loss: 0.005 [184, 360] loss: 0.005 Epoch: 184 -> Loss: 0.00239192019217 Epoch: 184 -> Test Accuracy: 92.27 [185, 60] loss: 0.005 [185, 120] loss: 0.005 [185, 180] loss: 0.005 [185, 240] loss: 0.006 [185, 300] loss: 0.006 [185, 360] loss: 0.004 Epoch: 185 -> Loss: 0.00132351741195 Epoch: 185 -> Test Accuracy: 92.18 [186, 60] loss: 0.004 [186, 120] loss: 0.004 [186, 180] loss: 0.005 [186, 240] loss: 0.005 [186, 300] loss: 0.005 [186, 360] loss: 0.006 Epoch: 186 -> Loss: 0.00507021322846 Epoch: 186 -> Test Accuracy: 92.2875 [187, 60] loss: 0.005 [187, 120] loss: 0.005 [187, 180] loss: 0.004 [187, 240] loss: 0.006 [187, 300] loss: 0.004 [187, 360] loss: 0.005 Epoch: 187 -> Loss: 0.0131164435297 Epoch: 187 -> Test Accuracy: 92.25 [188, 60] loss: 0.005 [188, 120] loss: 0.005 [188, 180] loss: 0.005 [188, 240] loss: 0.005 [188, 300] loss: 0.005 [188, 360] loss: 0.004 Epoch: 188 -> Loss: 0.00160925020464 Epoch: 188 -> Test Accuracy: 92.2125 [189, 60] loss: 0.004 [189, 120] loss: 0.004 [189, 180] loss: 0.004 [189, 240] loss: 0.005 [189, 300] loss: 0.005 [189, 360] loss: 0.005 Epoch: 189 -> Loss: 0.00413393136114 Epoch: 189 -> Test Accuracy: 92.175 [190, 60] loss: 0.005 [190, 120] loss: 0.005 [190, 180] loss: 0.005 [190, 240] loss: 0.005 [190, 300] loss: 0.005 [190, 360] loss: 0.005 Epoch: 190 -> Loss: 0.00773674622178 Epoch: 190 -> Test Accuracy: 92.195 [191, 60] loss: 0.004 [191, 120] loss: 0.004 [191, 180] loss: 0.005 [191, 240] loss: 0.004 [191, 300] loss: 0.004 [191, 360] loss: 0.004 Epoch: 191 -> Loss: 0.00615400727838 Epoch: 191 -> Test Accuracy: 92.2125 [192, 60] loss: 0.005 [192, 120] loss: 0.005 [192, 180] loss: 0.005 [192, 240] loss: 0.004 [192, 300] loss: 0.005 [192, 360] loss: 0.005 Epoch: 192 -> Loss: 0.00324656441808 Epoch: 192 -> Test Accuracy: 92.285 [193, 60] loss: 0.004 [193, 120] loss: 0.004 [193, 180] loss: 0.005 [193, 240] loss: 0.006 [193, 300] loss: 0.005 [193, 360] loss: 0.004 Epoch: 193 -> Loss: 0.00224489858374 Epoch: 193 -> Test Accuracy: 92.225 [194, 60] loss: 0.004 [194, 120] loss: 0.005 [194, 180] loss: 0.005 [194, 240] loss: 0.005 [194, 300] loss: 0.005 [194, 360] loss: 0.004 Epoch: 194 -> Loss: 0.00926404912025 Epoch: 194 -> Test Accuracy: 92.265 [195, 60] loss: 0.005 [195, 120] loss: 0.004 [195, 180] loss: 0.005 [195, 240] loss: 0.004 [195, 300] loss: 0.005 [195, 360] loss: 0.005 Epoch: 195 -> Loss: 0.00551159540191 Epoch: 195 -> Test Accuracy: 92.165 [196, 60] loss: 0.004 [196, 120] loss: 0.005 [196, 180] loss: 0.004 [196, 240] loss: 0.005 [196, 300] loss: 0.005 [196, 360] loss: 0.005 Epoch: 196 -> Loss: 0.00217827851884 Epoch: 196 -> Test Accuracy: 92.26 [197, 60] loss: 0.004 [197, 120] loss: 0.005 [197, 180] loss: 0.005 [197, 240] loss: 0.004 [197, 300] loss: 0.004 [197, 360] loss: 0.004 Epoch: 197 -> Loss: 0.00390889961272 Epoch: 197 -> Test Accuracy: 92.23 [198, 60] loss: 0.004 [198, 120] loss: 0.005 [198, 180] loss: 0.004 [198, 240] loss: 0.005 [198, 300] loss: 0.005 [198, 360] loss: 0.004 Epoch: 198 -> Loss: 0.0150258438662 Epoch: 198 -> Test Accuracy: 92.2525 [199, 60] loss: 0.004 [199, 120] loss: 0.005 [199, 180] loss: 0.004 [199, 240] loss: 0.005 [199, 300] loss: 0.004 [199, 360] loss: 0.004 Epoch: 199 -> Loss: 0.00687549123541 Epoch: 199 -> Test Accuracy: 92.2475 [200, 60] loss: 0.004 [200, 120] loss: 0.005 [200, 180] loss: 0.005 [200, 240] loss: 0.004 [200, 300] loss: 0.004 [200, 360] loss: 0.004 Epoch: 200 -> Loss: 0.00446217041463 Epoch: 200 -> Test Accuracy: 92.225 Finished Training
In [8]:
# train NonLinearClassifiers on feature map of net_3block
block5_loss_log, _, block5_test_accuracy_log, _, _ = tr.train_all_blocks(5, 10, [0.1, 0.02, 0.004, 0.0008],
[20, 40, 45, 100], 0.9, 5e-4, net_block5, criterion, trainloader, None, testloader)
[1, 60] loss: 2.206 [1, 120] loss: 1.259 [1, 180] loss: 1.152 [1, 240] loss: 1.113 [1, 300] loss: 1.040 [1, 360] loss: 1.005 Epoch: 1 -> Loss: 0.989844799042 Epoch: 1 -> Test Accuracy: 66.68 [2, 60] loss: 0.940 [2, 120] loss: 0.934 [2, 180] loss: 0.915 [2, 240] loss: 0.924 [2, 300] loss: 0.895 [2, 360] loss: 0.877 Epoch: 2 -> Loss: 0.998141288757 Epoch: 2 -> Test Accuracy: 71.31 [3, 60] loss: 0.845 [3, 120] loss: 0.839 [3, 180] loss: 0.833 [3, 240] loss: 0.820 [3, 300] loss: 0.793 [3, 360] loss: 0.805 Epoch: 3 -> Loss: 1.01133608818 Epoch: 3 -> Test Accuracy: 72.87 [4, 60] loss: 0.778 [4, 120] loss: 0.788 [4, 180] loss: 0.780 [4, 240] loss: 0.791 [4, 300] loss: 0.778 [4, 360] loss: 0.751 Epoch: 4 -> Loss: 0.792004227638 Epoch: 4 -> Test Accuracy: 73.82 [5, 60] loss: 0.761 [5, 120] loss: 0.749 [5, 180] loss: 0.750 [5, 240] loss: 0.741 [5, 300] loss: 0.738 [5, 360] loss: 0.718 Epoch: 5 -> Loss: 0.875502228737 Epoch: 5 -> Test Accuracy: 75.2 [6, 60] loss: 0.703 [6, 120] loss: 0.722 [6, 180] loss: 0.706 [6, 240] loss: 0.724 [6, 300] loss: 0.739 [6, 360] loss: 0.707 Epoch: 6 -> Loss: 0.882703483105 Epoch: 6 -> Test Accuracy: 75.69 [7, 60] loss: 0.675 [7, 120] loss: 0.677 [7, 180] loss: 0.704 [7, 240] loss: 0.685 [7, 300] loss: 0.726 [7, 360] loss: 0.681 Epoch: 7 -> Loss: 0.826486110687 Epoch: 7 -> Test Accuracy: 76.09 [8, 60] loss: 0.669 [8, 120] loss: 0.690 [8, 180] loss: 0.668 [8, 240] loss: 0.692 [8, 300] loss: 0.680 [8, 360] loss: 0.707 Epoch: 8 -> Loss: 0.748009562492 Epoch: 8 -> Test Accuracy: 76.21 [9, 60] loss: 0.679 [9, 120] loss: 0.666 [9, 180] loss: 0.657 [9, 240] loss: 0.680 [9, 300] loss: 0.684 [9, 360] loss: 0.688 Epoch: 9 -> Loss: 0.712456583977 Epoch: 9 -> Test Accuracy: 76.75 [10, 60] loss: 0.639 [10, 120] loss: 0.660 [10, 180] loss: 0.652 [10, 240] loss: 0.677 [10, 300] loss: 0.651 [10, 360] loss: 0.679 Epoch: 10 -> Loss: 0.672299444675 Epoch: 10 -> Test Accuracy: 77.05 [11, 60] loss: 0.652 [11, 120] loss: 0.665 [11, 180] loss: 0.662 [11, 240] loss: 0.656 [11, 300] loss: 0.642 [11, 360] loss: 0.661 Epoch: 11 -> Loss: 0.651005268097 Epoch: 11 -> Test Accuracy: 77.09 [12, 60] loss: 0.643 [12, 120] loss: 0.640 [12, 180] loss: 0.648 [12, 240] loss: 0.661 [12, 300] loss: 0.644 [12, 360] loss: 0.644 Epoch: 12 -> Loss: 0.640797972679 Epoch: 12 -> Test Accuracy: 76.95 [13, 60] loss: 0.635 [13, 120] loss: 0.621 [13, 180] loss: 0.653 [13, 240] loss: 0.632 [13, 300] loss: 0.644 [13, 360] loss: 0.651 Epoch: 13 -> Loss: 0.503665864468 Epoch: 13 -> Test Accuracy: 77.29 [14, 60] loss: 0.628 [14, 120] loss: 0.637 [14, 180] loss: 0.661 [14, 240] loss: 0.631 [14, 300] loss: 0.651 [14, 360] loss: 0.654 Epoch: 14 -> Loss: 0.545984447002 Epoch: 14 -> Test Accuracy: 77.82 [15, 60] loss: 0.595 [15, 120] loss: 0.633 [15, 180] loss: 0.631 [15, 240] loss: 0.650 [15, 300] loss: 0.637 [15, 360] loss: 0.642 Epoch: 15 -> Loss: 0.540837943554 Epoch: 15 -> Test Accuracy: 77.17 [16, 60] loss: 0.609 [16, 120] loss: 0.637 [16, 180] loss: 0.632 [16, 240] loss: 0.634 [16, 300] loss: 0.634 [16, 360] loss: 0.639 Epoch: 16 -> Loss: 0.786884486675 Epoch: 16 -> Test Accuracy: 77.27 [17, 60] loss: 0.599 [17, 120] loss: 0.627 [17, 180] loss: 0.609 [17, 240] loss: 0.634 [17, 300] loss: 0.649 [17, 360] loss: 0.640 Epoch: 17 -> Loss: 0.655207037926 Epoch: 17 -> Test Accuracy: 78.05 [18, 60] loss: 0.621 [18, 120] loss: 0.629 [18, 180] loss: 0.619 [18, 240] loss: 0.626 [18, 300] loss: 0.620 [18, 360] loss: 0.634 Epoch: 18 -> Loss: 0.673180103302 Epoch: 18 -> Test Accuracy: 77.42 [19, 60] loss: 0.612 [19, 120] loss: 0.615 [19, 180] loss: 0.627 [19, 240] loss: 0.600 [19, 300] loss: 0.629 [19, 360] loss: 0.620 Epoch: 19 -> Loss: 0.856933951378 Epoch: 19 -> Test Accuracy: 77.78 [20, 60] loss: 0.615 [20, 120] loss: 0.616 [20, 180] loss: 0.614 [20, 240] loss: 0.618 [20, 300] loss: 0.624 [20, 360] loss: 0.638 Epoch: 20 -> Loss: 0.642349123955 Epoch: 20 -> Test Accuracy: 77.8 [21, 60] loss: 0.555 [21, 120] loss: 0.550 [21, 180] loss: 0.512 [21, 240] loss: 0.516 [21, 300] loss: 0.510 [21, 360] loss: 0.518 Epoch: 21 -> Loss: 0.472536504269 Epoch: 21 -> Test Accuracy: 80.23 [22, 60] loss: 0.479 [22, 120] loss: 0.500 [22, 180] loss: 0.479 [22, 240] loss: 0.474 [22, 300] loss: 0.465 [22, 360] loss: 0.484 Epoch: 22 -> Loss: 0.399843990803 Epoch: 22 -> Test Accuracy: 80.83 [23, 60] loss: 0.482 [23, 120] loss: 0.454 [23, 180] loss: 0.453 [23, 240] loss: 0.465 [23, 300] loss: 0.465 [23, 360] loss: 0.461 Epoch: 23 -> Loss: 0.571590662003 Epoch: 23 -> Test Accuracy: 80.95 [24, 60] loss: 0.448 [24, 120] loss: 0.434 [24, 180] loss: 0.455 [24, 240] loss: 0.460 [24, 300] loss: 0.478 [24, 360] loss: 0.456 Epoch: 24 -> Loss: 0.458548158407 Epoch: 24 -> Test Accuracy: 81.19 [25, 60] loss: 0.441 [25, 120] loss: 0.441 [25, 180] loss: 0.447 [25, 240] loss: 0.449 [25, 300] loss: 0.429 [25, 360] loss: 0.461 Epoch: 25 -> Loss: 0.461026012897 Epoch: 25 -> Test Accuracy: 81.26 [26, 60] loss: 0.434 [26, 120] loss: 0.428 [26, 180] loss: 0.441 [26, 240] loss: 0.445 [26, 300] loss: 0.441 [26, 360] loss: 0.444 Epoch: 26 -> Loss: 0.455417454243 Epoch: 26 -> Test Accuracy: 81.19 [27, 60] loss: 0.427 [27, 120] loss: 0.425 [27, 180] loss: 0.449 [27, 240] loss: 0.437 [27, 300] loss: 0.447 [27, 360] loss: 0.423 Epoch: 27 -> Loss: 0.410543859005 Epoch: 27 -> Test Accuracy: 81.97 [28, 60] loss: 0.415 [28, 120] loss: 0.419 [28, 180] loss: 0.420 [28, 240] loss: 0.427 [28, 300] loss: 0.438 [28, 360] loss: 0.443 Epoch: 28 -> Loss: 0.532772958279 Epoch: 28 -> Test Accuracy: 81.37 [29, 60] loss: 0.418 [29, 120] loss: 0.407 [29, 180] loss: 0.406 [29, 240] loss: 0.414 [29, 300] loss: 0.427 [29, 360] loss: 0.426 Epoch: 29 -> Loss: 0.433989435434 Epoch: 29 -> Test Accuracy: 81.3 [30, 60] loss: 0.421 [30, 120] loss: 0.411 [30, 180] loss: 0.405 [30, 240] loss: 0.435 [30, 300] loss: 0.411 [30, 360] loss: 0.434 Epoch: 30 -> Loss: 0.345436096191 Epoch: 30 -> Test Accuracy: 81.55 [31, 60] loss: 0.396 [31, 120] loss: 0.426 [31, 180] loss: 0.427 [31, 240] loss: 0.425 [31, 300] loss: 0.418 [31, 360] loss: 0.418 Epoch: 31 -> Loss: 0.379749000072 Epoch: 31 -> Test Accuracy: 80.96 [32, 60] loss: 0.414 [32, 120] loss: 0.400 [32, 180] loss: 0.406 [32, 240] loss: 0.426 [32, 300] loss: 0.430 [32, 360] loss: 0.429 Epoch: 32 -> Loss: 0.340396940708 Epoch: 32 -> Test Accuracy: 80.91 [33, 60] loss: 0.405 [33, 120] loss: 0.415 [33, 180] loss: 0.428 [33, 240] loss: 0.408 [33, 300] loss: 0.405 [33, 360] loss: 0.421 Epoch: 33 -> Loss: 0.339397251606 Epoch: 33 -> Test Accuracy: 81.32 [34, 60] loss: 0.397 [34, 120] loss: 0.406 [34, 180] loss: 0.411 [34, 240] loss: 0.419 [34, 300] loss: 0.424 [34, 360] loss: 0.414 Epoch: 34 -> Loss: 0.495242774487 Epoch: 34 -> Test Accuracy: 81.44 [35, 60] loss: 0.380 [35, 120] loss: 0.419 [35, 180] loss: 0.411 [35, 240] loss: 0.409 [35, 300] loss: 0.420 [35, 360] loss: 0.406 Epoch: 35 -> Loss: 0.382247358561 Epoch: 35 -> Test Accuracy: 81.48 [36, 60] loss: 0.409 [36, 120] loss: 0.420 [36, 180] loss: 0.419 [36, 240] loss: 0.398 [36, 300] loss: 0.419 [36, 360] loss: 0.429 Epoch: 36 -> Loss: 0.416755497456 Epoch: 36 -> Test Accuracy: 81.15 [37, 60] loss: 0.407 [37, 120] loss: 0.409 [37, 180] loss: 0.402 [37, 240] loss: 0.401 [37, 300] loss: 0.401 [37, 360] loss: 0.433 Epoch: 37 -> Loss: 0.355153858662 Epoch: 37 -> Test Accuracy: 80.79 [38, 60] loss: 0.402 [38, 120] loss: 0.395 [38, 180] loss: 0.401 [38, 240] loss: 0.403 [38, 300] loss: 0.405 [38, 360] loss: 0.410 Epoch: 38 -> Loss: 0.381116777658 Epoch: 38 -> Test Accuracy: 80.63 [39, 60] loss: 0.398 [39, 120] loss: 0.408 [39, 180] loss: 0.409 [39, 240] loss: 0.397 [39, 300] loss: 0.409 [39, 360] loss: 0.438 Epoch: 39 -> Loss: 0.557063698769 Epoch: 39 -> Test Accuracy: 80.85 [40, 60] loss: 0.396 [40, 120] loss: 0.403 [40, 180] loss: 0.398 [40, 240] loss: 0.407 [40, 300] loss: 0.424 [40, 360] loss: 0.421 Epoch: 40 -> Loss: 0.344882249832 Epoch: 40 -> Test Accuracy: 81.3 [41, 60] loss: 0.384 [41, 120] loss: 0.367 [41, 180] loss: 0.370 [41, 240] loss: 0.342 [41, 300] loss: 0.361 [41, 360] loss: 0.352 Epoch: 41 -> Loss: 0.472389042377 Epoch: 41 -> Test Accuracy: 82.12 [42, 60] loss: 0.338 [42, 120] loss: 0.332 [42, 180] loss: 0.332 [42, 240] loss: 0.342 [42, 300] loss: 0.346 [42, 360] loss: 0.336 Epoch: 42 -> Loss: 0.429946184158 Epoch: 42 -> Test Accuracy: 82.26 [43, 60] loss: 0.325 [43, 120] loss: 0.326 [43, 180] loss: 0.332 [43, 240] loss: 0.324 [43, 300] loss: 0.325 [43, 360] loss: 0.335 Epoch: 43 -> Loss: 0.315294861794 Epoch: 43 -> Test Accuracy: 82.13 [44, 60] loss: 0.317 [44, 120] loss: 0.343 [44, 180] loss: 0.316 [44, 240] loss: 0.322 [44, 300] loss: 0.321 [44, 360] loss: 0.320 Epoch: 44 -> Loss: 0.436052948236 Epoch: 44 -> Test Accuracy: 82.21 [45, 60] loss: 0.323 [45, 120] loss: 0.308 [45, 180] loss: 0.323 [45, 240] loss: 0.324 [45, 300] loss: 0.307 [45, 360] loss: 0.319 Epoch: 45 -> Loss: 0.373412132263 Epoch: 45 -> Test Accuracy: 82.3 [46, 60] loss: 0.299 [46, 120] loss: 0.298 [46, 180] loss: 0.313 [46, 240] loss: 0.297 [46, 300] loss: 0.296 [46, 360] loss: 0.312 Epoch: 46 -> Loss: 0.242745012045 Epoch: 46 -> Test Accuracy: 82.47 [47, 60] loss: 0.293 [47, 120] loss: 0.297 [47, 180] loss: 0.295 [47, 240] loss: 0.305 [47, 300] loss: 0.301 [47, 360] loss: 0.304 Epoch: 47 -> Loss: 0.337700366974 Epoch: 47 -> Test Accuracy: 82.36 [48, 60] loss: 0.295 [48, 120] loss: 0.297 [48, 180] loss: 0.295 [48, 240] loss: 0.297 [48, 300] loss: 0.291 [48, 360] loss: 0.293 Epoch: 48 -> Loss: 0.408868640661 Epoch: 48 -> Test Accuracy: 82.4 [49, 60] loss: 0.294 [49, 120] loss: 0.305 [49, 180] loss: 0.293 [49, 240] loss: 0.301 [49, 300] loss: 0.294 [49, 360] loss: 0.288 Epoch: 49 -> Loss: 0.565850496292 Epoch: 49 -> Test Accuracy: 82.51 [50, 60] loss: 0.298 [50, 120] loss: 0.282 [50, 180] loss: 0.291 [50, 240] loss: 0.299 [50, 300] loss: 0.305 [50, 360] loss: 0.296 Epoch: 50 -> Loss: 0.328209519386 Epoch: 50 -> Test Accuracy: 82.43 [51, 60] loss: 0.279 [51, 120] loss: 0.292 [51, 180] loss: 0.298 [51, 240] loss: 0.283 [51, 300] loss: 0.287 [51, 360] loss: 0.300 Epoch: 51 -> Loss: 0.325222551823 Epoch: 51 -> Test Accuracy: 82.38 [52, 60] loss: 0.282 [52, 120] loss: 0.280 [52, 180] loss: 0.278 [52, 240] loss: 0.294 [52, 300] loss: 0.302 [52, 360] loss: 0.304 Epoch: 52 -> Loss: 0.353080123663 Epoch: 52 -> Test Accuracy: 82.46 [53, 60] loss: 0.274 [53, 120] loss: 0.287 [53, 180] loss: 0.280 [53, 240] loss: 0.304 [53, 300] loss: 0.289 [53, 360] loss: 0.277 Epoch: 53 -> Loss: 0.293329536915 Epoch: 53 -> Test Accuracy: 82.55 [54, 60] loss: 0.294 [54, 120] loss: 0.284 [54, 180] loss: 0.278 [54, 240] loss: 0.285 [54, 300] loss: 0.288 [54, 360] loss: 0.298 Epoch: 54 -> Loss: 0.22264918685 Epoch: 54 -> Test Accuracy: 82.44 [55, 60] loss: 0.296 [55, 120] loss: 0.291 [55, 180] loss: 0.287 [55, 240] loss: 0.289 [55, 300] loss: 0.275 [55, 360] loss: 0.275 Epoch: 55 -> Loss: 0.552338838577 Epoch: 55 -> Test Accuracy: 82.44 [56, 60] loss: 0.286 [56, 120] loss: 0.281 [56, 180] loss: 0.280 [56, 240] loss: 0.284 [56, 300] loss: 0.286 [56, 360] loss: 0.279 Epoch: 56 -> Loss: 0.26289281249 Epoch: 56 -> Test Accuracy: 82.51 [57, 60] loss: 0.270 [57, 120] loss: 0.276 [57, 180] loss: 0.276 [57, 240] loss: 0.287 [57, 300] loss: 0.291 [57, 360] loss: 0.282 Epoch: 57 -> Loss: 0.189671799541 Epoch: 57 -> Test Accuracy: 82.6 [58, 60] loss: 0.290 [58, 120] loss: 0.264 [58, 180] loss: 0.285 [58, 240] loss: 0.287 [58, 300] loss: 0.290 [58, 360] loss: 0.278 Epoch: 58 -> Loss: 0.243798166513 Epoch: 58 -> Test Accuracy: 82.53 [59, 60] loss: 0.279 [59, 120] loss: 0.285 [59, 180] loss: 0.275 [59, 240] loss: 0.289 [59, 300] loss: 0.287 [59, 360] loss: 0.283 Epoch: 59 -> Loss: 0.253309726715 Epoch: 59 -> Test Accuracy: 82.5 [60, 60] loss: 0.287 [60, 120] loss: 0.279 [60, 180] loss: 0.272 [60, 240] loss: 0.277 [60, 300] loss: 0.269 [60, 360] loss: 0.270 Epoch: 60 -> Loss: 0.283162623644 Epoch: 60 -> Test Accuracy: 82.51 [61, 60] loss: 0.274 [61, 120] loss: 0.273 [61, 180] loss: 0.286 [61, 240] loss: 0.275 [61, 300] loss: 0.272 [61, 360] loss: 0.285 Epoch: 61 -> Loss: 0.313937842846 Epoch: 61 -> Test Accuracy: 82.63 [62, 60] loss: 0.283 [62, 120] loss: 0.274 [62, 180] loss: 0.262 [62, 240] loss: 0.280 [62, 300] loss: 0.275 [62, 360] loss: 0.283 Epoch: 62 -> Loss: 0.335236012936 Epoch: 62 -> Test Accuracy: 82.52 [63, 60] loss: 0.285 [63, 120] loss: 0.277 [63, 180] loss: 0.279 [63, 240] loss: 0.264 [63, 300] loss: 0.277 [63, 360] loss: 0.278 Epoch: 63 -> Loss: 0.311462879181 Epoch: 63 -> Test Accuracy: 82.52 [64, 60] loss: 0.271 [64, 120] loss: 0.272 [64, 180] loss: 0.273 [64, 240] loss: 0.268 [64, 300] loss: 0.290 [64, 360] loss: 0.279 Epoch: 64 -> Loss: 0.222185850143 Epoch: 64 -> Test Accuracy: 82.72 [65, 60] loss: 0.269 [65, 120] loss: 0.286 [65, 180] loss: 0.276 [65, 240] loss: 0.285 [65, 300] loss: 0.275 [65, 360] loss: 0.262 Epoch: 65 -> Loss: 0.18132416904 Epoch: 65 -> Test Accuracy: 82.74 [66, 60] loss: 0.266 [66, 120] loss: 0.272 [66, 180] loss: 0.267 [66, 240] loss: 0.273 [66, 300] loss: 0.279 [66, 360] loss: 0.265 Epoch: 66 -> Loss: 0.226240590215 Epoch: 66 -> Test Accuracy: 82.53 [67, 60] loss: 0.276 [67, 120] loss: 0.251 [67, 180] loss: 0.271 [67, 240] loss: 0.262 [67, 300] loss: 0.277 [67, 360] loss: 0.269 Epoch: 67 -> Loss: 0.19538384676 Epoch: 67 -> Test Accuracy: 82.62 [68, 60] loss: 0.276 [68, 120] loss: 0.273 [68, 180] loss: 0.271 [68, 240] loss: 0.270 [68, 300] loss: 0.268 [68, 360] loss: 0.263 Epoch: 68 -> Loss: 0.344725430012 Epoch: 68 -> Test Accuracy: 82.66 [69, 60] loss: 0.270 [69, 120] loss: 0.260 [69, 180] loss: 0.274 [69, 240] loss: 0.269 [69, 300] loss: 0.264 [69, 360] loss: 0.258 Epoch: 69 -> Loss: 0.245025873184 Epoch: 69 -> Test Accuracy: 82.71 [70, 60] loss: 0.271 [70, 120] loss: 0.267 [70, 180] loss: 0.269 [70, 240] loss: 0.281 [70, 300] loss: 0.257 [70, 360] loss: 0.275 Epoch: 70 -> Loss: 0.29664465785 Epoch: 70 -> Test Accuracy: 82.57 [71, 60] loss: 0.265 [71, 120] loss: 0.266 [71, 180] loss: 0.264 [71, 240] loss: 0.262 [71, 300] loss: 0.262 [71, 360] loss: 0.273 Epoch: 71 -> Loss: 0.216807082295 Epoch: 71 -> Test Accuracy: 82.67 [72, 60] loss: 0.264 [72, 120] loss: 0.267 [72, 180] loss: 0.264 [72, 240] loss: 0.261 [72, 300] loss: 0.274 [72, 360] loss: 0.274 Epoch: 72 -> Loss: 0.3693177104 Epoch: 72 -> Test Accuracy: 82.4 [73, 60] loss: 0.275 [73, 120] loss: 0.255 [73, 180] loss: 0.269 [73, 240] loss: 0.264 [73, 300] loss: 0.257 [73, 360] loss: 0.270 Epoch: 73 -> Loss: 0.250025957823 Epoch: 73 -> Test Accuracy: 82.48 [74, 60] loss: 0.260 [74, 120] loss: 0.273 [74, 180] loss: 0.266 [74, 240] loss: 0.258 [74, 300] loss: 0.261 [74, 360] loss: 0.263 Epoch: 74 -> Loss: 0.241463631392 Epoch: 74 -> Test Accuracy: 82.53 [75, 60] loss: 0.257 [75, 120] loss: 0.263 [75, 180] loss: 0.270 [75, 240] loss: 0.270 [75, 300] loss: 0.253 [75, 360] loss: 0.265 Epoch: 75 -> Loss: 0.297407001257 Epoch: 75 -> Test Accuracy: 82.38 [76, 60] loss: 0.245 [76, 120] loss: 0.259 [76, 180] loss: 0.278 [76, 240] loss: 0.260 [76, 300] loss: 0.276 [76, 360] loss: 0.273 Epoch: 76 -> Loss: 0.329516738653 Epoch: 76 -> Test Accuracy: 82.29 [77, 60] loss: 0.257 [77, 120] loss: 0.256 [77, 180] loss: 0.270 [77, 240] loss: 0.257 [77, 300] loss: 0.258 [77, 360] loss: 0.260 Epoch: 77 -> Loss: 0.222680807114 Epoch: 77 -> Test Accuracy: 82.39 [78, 60] loss: 0.262 [78, 120] loss: 0.265 [78, 180] loss: 0.267 [78, 240] loss: 0.254 [78, 300] loss: 0.261 [78, 360] loss: 0.263 Epoch: 78 -> Loss: 0.245288655162 Epoch: 78 -> Test Accuracy: 82.49 [79, 60] loss: 0.257 [79, 120] loss: 0.259 [79, 180] loss: 0.259 [79, 240] loss: 0.254 [79, 300] loss: 0.266 [79, 360] loss: 0.275 Epoch: 79 -> Loss: 0.277887523174 Epoch: 79 -> Test Accuracy: 82.54 [80, 60] loss: 0.260 [80, 120] loss: 0.261 [80, 180] loss: 0.260 [80, 240] loss: 0.260 [80, 300] loss: 0.261 [80, 360] loss: 0.258 Epoch: 80 -> Loss: 0.214941501617 Epoch: 80 -> Test Accuracy: 82.62 [81, 60] loss: 0.264 [81, 120] loss: 0.264 [81, 180] loss: 0.259 [81, 240] loss: 0.257 [81, 300] loss: 0.254 [81, 360] loss: 0.267 Epoch: 81 -> Loss: 0.219567447901 Epoch: 81 -> Test Accuracy: 82.65 [82, 60] loss: 0.259 [82, 120] loss: 0.254 [82, 180] loss: 0.255 [82, 240] loss: 0.257 [82, 300] loss: 0.265 [82, 360] loss: 0.253 Epoch: 82 -> Loss: 0.262077003717 Epoch: 82 -> Test Accuracy: 82.56 [83, 60] loss: 0.254 [83, 120] loss: 0.270 [83, 180] loss: 0.262 [83, 240] loss: 0.248 [83, 300] loss: 0.262 [83, 360] loss: 0.255 Epoch: 83 -> Loss: 0.187527015805 Epoch: 83 -> Test Accuracy: 82.52 [84, 60] loss: 0.244 [84, 120] loss: 0.251 [84, 180] loss: 0.251 [84, 240] loss: 0.259 [84, 300] loss: 0.267 [84, 360] loss: 0.263 Epoch: 84 -> Loss: 0.235802292824 Epoch: 84 -> Test Accuracy: 82.6 [85, 60] loss: 0.251 [85, 120] loss: 0.248 [85, 180] loss: 0.251 [85, 240] loss: 0.255 [85, 300] loss: 0.259 [85, 360] loss: 0.244 Epoch: 85 -> Loss: 0.219255179167 Epoch: 85 -> Test Accuracy: 82.75 [86, 60] loss: 0.257 [86, 120] loss: 0.253 [86, 180] loss: 0.261 [86, 240] loss: 0.248 [86, 300] loss: 0.247 [86, 360] loss: 0.259 Epoch: 86 -> Loss: 0.169390812516 Epoch: 86 -> Test Accuracy: 82.82 [87, 60] loss: 0.267 [87, 120] loss: 0.256 [87, 180] loss: 0.249 [87, 240] loss: 0.249 [87, 300] loss: 0.252 [87, 360] loss: 0.252 Epoch: 87 -> Loss: 0.29005163908 Epoch: 87 -> Test Accuracy: 82.6 [88, 60] loss: 0.257 [88, 120] loss: 0.257 [88, 180] loss: 0.247 [88, 240] loss: 0.243 [88, 300] loss: 0.249 [88, 360] loss: 0.254 Epoch: 88 -> Loss: 0.368238866329 Epoch: 88 -> Test Accuracy: 82.58 [89, 60] loss: 0.253 [89, 120] loss: 0.267 [89, 180] loss: 0.243 [89, 240] loss: 0.254 [89, 300] loss: 0.256 [89, 360] loss: 0.253 Epoch: 89 -> Loss: 0.257168114185 Epoch: 89 -> Test Accuracy: 82.64 [90, 60] loss: 0.254 [90, 120] loss: 0.251 [90, 180] loss: 0.239 [90, 240] loss: 0.258 [90, 300] loss: 0.249 [90, 360] loss: 0.259 Epoch: 90 -> Loss: 0.238972112536 Epoch: 90 -> Test Accuracy: 82.66 [91, 60] loss: 0.243 [91, 120] loss: 0.253 [91, 180] loss: 0.246 [91, 240] loss: 0.253 [91, 300] loss: 0.256 [91, 360] loss: 0.249 Epoch: 91 -> Loss: 0.373318344355 Epoch: 91 -> Test Accuracy: 82.67 [92, 60] loss: 0.247 [92, 120] loss: 0.256 [92, 180] loss: 0.244 [92, 240] loss: 0.251 [92, 300] loss: 0.247 [92, 360] loss: 0.250 Epoch: 92 -> Loss: 0.139528647065 Epoch: 92 -> Test Accuracy: 82.53 [93, 60] loss: 0.244 [93, 120] loss: 0.238 [93, 180] loss: 0.250 [93, 240] loss: 0.251 [93, 300] loss: 0.252 [93, 360] loss: 0.245 Epoch: 93 -> Loss: 0.251868247986 Epoch: 93 -> Test Accuracy: 82.57 [94, 60] loss: 0.244 [94, 120] loss: 0.251 [94, 180] loss: 0.242 [94, 240] loss: 0.243 [94, 300] loss: 0.244 [94, 360] loss: 0.264 Epoch: 94 -> Loss: 0.244362637401 Epoch: 94 -> Test Accuracy: 82.45 [95, 60] loss: 0.244 [95, 120] loss: 0.252 [95, 180] loss: 0.238 [95, 240] loss: 0.241 [95, 300] loss: 0.248 [95, 360] loss: 0.255 Epoch: 95 -> Loss: 0.222351387143 Epoch: 95 -> Test Accuracy: 82.59 [96, 60] loss: 0.241 [96, 120] loss: 0.242 [96, 180] loss: 0.235 [96, 240] loss: 0.248 [96, 300] loss: 0.248 [96, 360] loss: 0.251 Epoch: 96 -> Loss: 0.273922264576 Epoch: 96 -> Test Accuracy: 82.48 [97, 60] loss: 0.239 [97, 120] loss: 0.250 [97, 180] loss: 0.255 [97, 240] loss: 0.241 [97, 300] loss: 0.244 [97, 360] loss: 0.246 Epoch: 97 -> Loss: 0.318357616663 Epoch: 97 -> Test Accuracy: 82.69 [98, 60] loss: 0.244 [98, 120] loss: 0.247 [98, 180] loss: 0.242 [98, 240] loss: 0.240 [98, 300] loss: 0.244 [98, 360] loss: 0.234 Epoch: 98 -> Loss: 0.209844306111 Epoch: 98 -> Test Accuracy: 82.58 [99, 60] loss: 0.233 [99, 120] loss: 0.234 [99, 180] loss: 0.233 [99, 240] loss: 0.253 [99, 300] loss: 0.259 [99, 360] loss: 0.249 Epoch: 99 -> Loss: 0.286715209484 Epoch: 99 -> Test Accuracy: 82.71 [100, 60] loss: 0.242 [100, 120] loss: 0.253 [100, 180] loss: 0.252 [100, 240] loss: 0.238 [100, 300] loss: 0.251 [100, 360] loss: 0.237 Epoch: 100 -> Loss: 0.240406125784 Epoch: 100 -> Test Accuracy: 82.64 Finished Training [1, 60] loss: 1.745 [1, 120] loss: 0.841 [1, 180] loss: 0.763 [1, 240] loss: 0.717 [1, 300] loss: 0.704 [1, 360] loss: 0.660 Epoch: 1 -> Loss: 0.522156357765 Epoch: 1 -> Test Accuracy: 77.76 [2, 60] loss: 0.598 [2, 120] loss: 0.610 [2, 180] loss: 0.584 [2, 240] loss: 0.577 [2, 300] loss: 0.562 [2, 360] loss: 0.563 Epoch: 2 -> Loss: 0.696018874645 Epoch: 2 -> Test Accuracy: 80.68 [3, 60] loss: 0.511 [3, 120] loss: 0.534 [3, 180] loss: 0.516 [3, 240] loss: 0.514 [3, 300] loss: 0.521 [3, 360] loss: 0.526 Epoch: 3 -> Loss: 0.558457434177 Epoch: 3 -> Test Accuracy: 81.52 [4, 60] loss: 0.475 [4, 120] loss: 0.470 [4, 180] loss: 0.496 [4, 240] loss: 0.487 [4, 300] loss: 0.489 [4, 360] loss: 0.489 Epoch: 4 -> Loss: 0.476736158133 Epoch: 4 -> Test Accuracy: 82.34 [5, 60] loss: 0.460 [5, 120] loss: 0.463 [5, 180] loss: 0.449 [5, 240] loss: 0.465 [5, 300] loss: 0.471 [5, 360] loss: 0.469 Epoch: 5 -> Loss: 0.391936689615 Epoch: 5 -> Test Accuracy: 82.78 [6, 60] loss: 0.422 [6, 120] loss: 0.439 [6, 180] loss: 0.446 [6, 240] loss: 0.453 [6, 300] loss: 0.440 [6, 360] loss: 0.438 Epoch: 6 -> Loss: 0.426587998867 Epoch: 6 -> Test Accuracy: 83.13 [7, 60] loss: 0.409 [7, 120] loss: 0.443 [7, 180] loss: 0.433 [7, 240] loss: 0.438 [7, 300] loss: 0.426 [7, 360] loss: 0.437 Epoch: 7 -> Loss: 0.472709596157 Epoch: 7 -> Test Accuracy: 83.3 [8, 60] loss: 0.402 [8, 120] loss: 0.413 [8, 180] loss: 0.437 [8, 240] loss: 0.416 [8, 300] loss: 0.433 [8, 360] loss: 0.432 Epoch: 8 -> Loss: 0.505204975605 Epoch: 8 -> Test Accuracy: 83.45 [9, 60] loss: 0.390 [9, 120] loss: 0.413 [9, 180] loss: 0.426 [9, 240] loss: 0.412 [9, 300] loss: 0.414 [9, 360] loss: 0.417 Epoch: 9 -> Loss: 0.393939316273 Epoch: 9 -> Test Accuracy: 82.76 [10, 60] loss: 0.412 [10, 120] loss: 0.396 [10, 180] loss: 0.405 [10, 240] loss: 0.398 [10, 300] loss: 0.427 [10, 360] loss: 0.421 Epoch: 10 -> Loss: 0.536645770073 Epoch: 10 -> Test Accuracy: 83.96 [11, 60] loss: 0.393 [11, 120] loss: 0.374 [11, 180] loss: 0.423 [11, 240] loss: 0.394 [11, 300] loss: 0.418 [11, 360] loss: 0.407 Epoch: 11 -> Loss: 0.675051510334 Epoch: 11 -> Test Accuracy: 83.22 [12, 60] loss: 0.370 [12, 120] loss: 0.392 [12, 180] loss: 0.391 [12, 240] loss: 0.402 [12, 300] loss: 0.399 [12, 360] loss: 0.399 Epoch: 12 -> Loss: 0.475282132626 Epoch: 12 -> Test Accuracy: 83.68 [13, 60] loss: 0.375 [13, 120] loss: 0.386 [13, 180] loss: 0.395 [13, 240] loss: 0.390 [13, 300] loss: 0.417 [13, 360] loss: 0.397 Epoch: 13 -> Loss: 0.470419168472 Epoch: 13 -> Test Accuracy: 83.8 [14, 60] loss: 0.367 [14, 120] loss: 0.373 [14, 180] loss: 0.387 [14, 240] loss: 0.399 [14, 300] loss: 0.400 [14, 360] loss: 0.403 Epoch: 14 -> Loss: 0.278404980898 Epoch: 14 -> Test Accuracy: 83.76 [15, 60] loss: 0.372 [15, 120] loss: 0.375 [15, 180] loss: 0.376 [15, 240] loss: 0.384 [15, 300] loss: 0.406 [15, 360] loss: 0.399 Epoch: 15 -> Loss: 0.458360135555 Epoch: 15 -> Test Accuracy: 83.84 [16, 60] loss: 0.350 [16, 120] loss: 0.382 [16, 180] loss: 0.404 [16, 240] loss: 0.370 [16, 300] loss: 0.387 [16, 360] loss: 0.388 Epoch: 16 -> Loss: 0.542204737663 Epoch: 16 -> Test Accuracy: 84.4 [17, 60] loss: 0.370 [17, 120] loss: 0.357 [17, 180] loss: 0.377 [17, 240] loss: 0.364 [17, 300] loss: 0.391 [17, 360] loss: 0.408 Epoch: 17 -> Loss: 0.406774371862 Epoch: 17 -> Test Accuracy: 83.69 [18, 60] loss: 0.361 [18, 120] loss: 0.370 [18, 180] loss: 0.374 [18, 240] loss: 0.381 [18, 300] loss: 0.382 [18, 360] loss: 0.378 Epoch: 18 -> Loss: 0.390609174967 Epoch: 18 -> Test Accuracy: 83.99 [19, 60] loss: 0.355 [19, 120] loss: 0.379 [19, 180] loss: 0.360 [19, 240] loss: 0.388 [19, 300] loss: 0.384 [19, 360] loss: 0.398 Epoch: 19 -> Loss: 0.531518101692 Epoch: 19 -> Test Accuracy: 84.16 [20, 60] loss: 0.353 [20, 120] loss: 0.364 [20, 180] loss: 0.369 [20, 240] loss: 0.381 [20, 300] loss: 0.390 [20, 360] loss: 0.372 Epoch: 20 -> Loss: 0.658749222755 Epoch: 20 -> Test Accuracy: 84.1 [21, 60] loss: 0.335 [21, 120] loss: 0.305 [21, 180] loss: 0.287 [21, 240] loss: 0.294 [21, 300] loss: 0.300 [21, 360] loss: 0.288 Epoch: 21 -> Loss: 0.237031057477 Epoch: 21 -> Test Accuracy: 86.04 [22, 60] loss: 0.256 [22, 120] loss: 0.286 [22, 180] loss: 0.250 [22, 240] loss: 0.273 [22, 300] loss: 0.269 [22, 360] loss: 0.267 Epoch: 22 -> Loss: 0.301119506359 Epoch: 22 -> Test Accuracy: 86.37 [23, 60] loss: 0.253 [23, 120] loss: 0.236 [23, 180] loss: 0.257 [23, 240] loss: 0.247 [23, 300] loss: 0.248 [23, 360] loss: 0.244 Epoch: 23 -> Loss: 0.214704036713 Epoch: 23 -> Test Accuracy: 86.59 [24, 60] loss: 0.240 [24, 120] loss: 0.239 [24, 180] loss: 0.237 [24, 240] loss: 0.246 [24, 300] loss: 0.238 [24, 360] loss: 0.242 Epoch: 24 -> Loss: 0.138725206256 Epoch: 24 -> Test Accuracy: 86.59 [25, 60] loss: 0.225 [25, 120] loss: 0.220 [25, 180] loss: 0.215 [25, 240] loss: 0.240 [25, 300] loss: 0.232 [25, 360] loss: 0.227 Epoch: 25 -> Loss: 0.226364284754 Epoch: 25 -> Test Accuracy: 86.27 [26, 60] loss: 0.204 [26, 120] loss: 0.211 [26, 180] loss: 0.230 [26, 240] loss: 0.226 [26, 300] loss: 0.216 [26, 360] loss: 0.224 Epoch: 26 -> Loss: 0.19846278429 Epoch: 26 -> Test Accuracy: 86.16 [27, 60] loss: 0.205 [27, 120] loss: 0.211 [27, 180] loss: 0.219 [27, 240] loss: 0.222 [27, 300] loss: 0.206 [27, 360] loss: 0.225 Epoch: 27 -> Loss: 0.245066836476 Epoch: 27 -> Test Accuracy: 86.59 [28, 60] loss: 0.205 [28, 120] loss: 0.201 [28, 180] loss: 0.215 [28, 240] loss: 0.210 [28, 300] loss: 0.225 [28, 360] loss: 0.207 Epoch: 28 -> Loss: 0.256059736013 Epoch: 28 -> Test Accuracy: 86.18 [29, 60] loss: 0.202 [29, 120] loss: 0.202 [29, 180] loss: 0.225 [29, 240] loss: 0.209 [29, 300] loss: 0.218 [29, 360] loss: 0.212 Epoch: 29 -> Loss: 0.21028470993 Epoch: 29 -> Test Accuracy: 86.18 [30, 60] loss: 0.201 [30, 120] loss: 0.205 [30, 180] loss: 0.201 [30, 240] loss: 0.205 [30, 300] loss: 0.205 [30, 360] loss: 0.215 Epoch: 30 -> Loss: 0.233492895961 Epoch: 30 -> Test Accuracy: 86.2 [31, 60] loss: 0.199 [31, 120] loss: 0.184 [31, 180] loss: 0.197 [31, 240] loss: 0.209 [31, 300] loss: 0.214 [31, 360] loss: 0.212 Epoch: 31 -> Loss: 0.157852530479 Epoch: 31 -> Test Accuracy: 85.71 [32, 60] loss: 0.200 [32, 120] loss: 0.211 [32, 180] loss: 0.197 [32, 240] loss: 0.196 [32, 300] loss: 0.223 [32, 360] loss: 0.212 Epoch: 32 -> Loss: 0.228305101395 Epoch: 32 -> Test Accuracy: 85.88 [33, 60] loss: 0.187 [33, 120] loss: 0.195 [33, 180] loss: 0.205 [33, 240] loss: 0.201 [33, 300] loss: 0.221 [33, 360] loss: 0.207 Epoch: 33 -> Loss: 0.162550657988 Epoch: 33 -> Test Accuracy: 85.6 [34, 60] loss: 0.185 [34, 120] loss: 0.194 [34, 180] loss: 0.190 [34, 240] loss: 0.203 [34, 300] loss: 0.216 [34, 360] loss: 0.214 Epoch: 34 -> Loss: 0.145446151495 Epoch: 34 -> Test Accuracy: 85.77 [35, 60] loss: 0.193 [35, 120] loss: 0.201 [35, 180] loss: 0.202 [35, 240] loss: 0.213 [35, 300] loss: 0.217 [35, 360] loss: 0.197 Epoch: 35 -> Loss: 0.267737209797 Epoch: 35 -> Test Accuracy: 86.16 [36, 60] loss: 0.197 [36, 120] loss: 0.200 [36, 180] loss: 0.198 [36, 240] loss: 0.203 [36, 300] loss: 0.201 [36, 360] loss: 0.209 Epoch: 36 -> Loss: 0.198297768831 Epoch: 36 -> Test Accuracy: 86.07 [37, 60] loss: 0.190 [37, 120] loss: 0.185 [37, 180] loss: 0.209 [37, 240] loss: 0.201 [37, 300] loss: 0.212 [37, 360] loss: 0.214 Epoch: 37 -> Loss: 0.162070557475 Epoch: 37 -> Test Accuracy: 85.59 [38, 60] loss: 0.193 [38, 120] loss: 0.194 [38, 180] loss: 0.203 [38, 240] loss: 0.193 [38, 300] loss: 0.198 [38, 360] loss: 0.213 Epoch: 38 -> Loss: 0.171396017075 Epoch: 38 -> Test Accuracy: 85.38 [39, 60] loss: 0.197 [39, 120] loss: 0.198 [39, 180] loss: 0.205 [39, 240] loss: 0.196 [39, 300] loss: 0.210 [39, 360] loss: 0.198 Epoch: 39 -> Loss: 0.184170261025 Epoch: 39 -> Test Accuracy: 85.57 [40, 60] loss: 0.184 [40, 120] loss: 0.182 [40, 180] loss: 0.185 [40, 240] loss: 0.199 [40, 300] loss: 0.207 [40, 360] loss: 0.203 Epoch: 40 -> Loss: 0.0648645684123 Epoch: 40 -> Test Accuracy: 85.76 [41, 60] loss: 0.178 [41, 120] loss: 0.179 [41, 180] loss: 0.173 [41, 240] loss: 0.157 [41, 300] loss: 0.156 [41, 360] loss: 0.154 Epoch: 41 -> Loss: 0.128126725554 Epoch: 41 -> Test Accuracy: 86.48 [42, 60] loss: 0.147 [42, 120] loss: 0.149 [42, 180] loss: 0.142 [42, 240] loss: 0.143 [42, 300] loss: 0.147 [42, 360] loss: 0.151 Epoch: 42 -> Loss: 0.118626393378 Epoch: 42 -> Test Accuracy: 86.84 [43, 60] loss: 0.137 [43, 120] loss: 0.145 [43, 180] loss: 0.141 [43, 240] loss: 0.136 [43, 300] loss: 0.142 [43, 360] loss: 0.135 Epoch: 43 -> Loss: 0.197865873575 Epoch: 43 -> Test Accuracy: 86.87 [44, 60] loss: 0.120 [44, 120] loss: 0.134 [44, 180] loss: 0.133 [44, 240] loss: 0.123 [44, 300] loss: 0.128 [44, 360] loss: 0.127 Epoch: 44 -> Loss: 0.156142085791 Epoch: 44 -> Test Accuracy: 86.55 [45, 60] loss: 0.121 [45, 120] loss: 0.127 [45, 180] loss: 0.136 [45, 240] loss: 0.128 [45, 300] loss: 0.134 [45, 360] loss: 0.129 Epoch: 45 -> Loss: 0.120640911162 Epoch: 45 -> Test Accuracy: 87.09 [46, 60] loss: 0.119 [46, 120] loss: 0.123 [46, 180] loss: 0.116 [46, 240] loss: 0.110 [46, 300] loss: 0.120 [46, 360] loss: 0.118 Epoch: 46 -> Loss: 0.119994021952 Epoch: 46 -> Test Accuracy: 87.12 [47, 60] loss: 0.118 [47, 120] loss: 0.116 [47, 180] loss: 0.115 [47, 240] loss: 0.126 [47, 300] loss: 0.119 [47, 360] loss: 0.115 Epoch: 47 -> Loss: 0.0461328849196 Epoch: 47 -> Test Accuracy: 87.06 [48, 60] loss: 0.119 [48, 120] loss: 0.119 [48, 180] loss: 0.120 [48, 240] loss: 0.111 [48, 300] loss: 0.115 [48, 360] loss: 0.121 Epoch: 48 -> Loss: 0.0995544865727 Epoch: 48 -> Test Accuracy: 87.16 [49, 60] loss: 0.119 [49, 120] loss: 0.114 [49, 180] loss: 0.110 [49, 240] loss: 0.118 [49, 300] loss: 0.116 [49, 360] loss: 0.107 Epoch: 49 -> Loss: 0.18508708477 Epoch: 49 -> Test Accuracy: 87.23 [50, 60] loss: 0.113 [50, 120] loss: 0.106 [50, 180] loss: 0.112 [50, 240] loss: 0.109 [50, 300] loss: 0.121 [50, 360] loss: 0.115 Epoch: 50 -> Loss: 0.225960090756 Epoch: 50 -> Test Accuracy: 87.3 [51, 60] loss: 0.111 [51, 120] loss: 0.113 [51, 180] loss: 0.113 [51, 240] loss: 0.111 [51, 300] loss: 0.109 [51, 360] loss: 0.104 Epoch: 51 -> Loss: 0.134554818273 Epoch: 51 -> Test Accuracy: 87.08 [52, 60] loss: 0.112 [52, 120] loss: 0.107 [52, 180] loss: 0.104 [52, 240] loss: 0.111 [52, 300] loss: 0.097 [52, 360] loss: 0.120 Epoch: 52 -> Loss: 0.11552156508 Epoch: 52 -> Test Accuracy: 87.08 [53, 60] loss: 0.104 [53, 120] loss: 0.115 [53, 180] loss: 0.110 [53, 240] loss: 0.112 [53, 300] loss: 0.101 [53, 360] loss: 0.106 Epoch: 53 -> Loss: 0.136014476418 Epoch: 53 -> Test Accuracy: 87.19 [54, 60] loss: 0.109 [54, 120] loss: 0.104 [54, 180] loss: 0.102 [54, 240] loss: 0.107 [54, 300] loss: 0.108 [54, 360] loss: 0.108 Epoch: 54 -> Loss: 0.127010077238 Epoch: 54 -> Test Accuracy: 87.15 [55, 60] loss: 0.106 [55, 120] loss: 0.095 [55, 180] loss: 0.107 [55, 240] loss: 0.107 [55, 300] loss: 0.104 [55, 360] loss: 0.106 Epoch: 55 -> Loss: 0.148679152131 Epoch: 55 -> Test Accuracy: 87.08 [56, 60] loss: 0.103 [56, 120] loss: 0.102 [56, 180] loss: 0.103 [56, 240] loss: 0.109 [56, 300] loss: 0.102 [56, 360] loss: 0.107 Epoch: 56 -> Loss: 0.116108573973 Epoch: 56 -> Test Accuracy: 87.21 [57, 60] loss: 0.109 [57, 120] loss: 0.105 [57, 180] loss: 0.103 [57, 240] loss: 0.103 [57, 300] loss: 0.112 [57, 360] loss: 0.107 Epoch: 57 -> Loss: 0.121085688472 Epoch: 57 -> Test Accuracy: 87.12 [58, 60] loss: 0.102 [58, 120] loss: 0.099 [58, 180] loss: 0.104 [58, 240] loss: 0.108 [58, 300] loss: 0.104 [58, 360] loss: 0.104 Epoch: 58 -> Loss: 0.150178059936 Epoch: 58 -> Test Accuracy: 87.28 [59, 60] loss: 0.099 [59, 120] loss: 0.098 [59, 180] loss: 0.101 [59, 240] loss: 0.110 [59, 300] loss: 0.110 [59, 360] loss: 0.106 Epoch: 59 -> Loss: 0.053861014545 Epoch: 59 -> Test Accuracy: 87.17 [60, 60] loss: 0.102 [60, 120] loss: 0.093 [60, 180] loss: 0.097 [60, 240] loss: 0.097 [60, 300] loss: 0.099 [60, 360] loss: 0.105 Epoch: 60 -> Loss: 0.18052020669 Epoch: 60 -> Test Accuracy: 87.17 [61, 60] loss: 0.103 [61, 120] loss: 0.098 [61, 180] loss: 0.097 [61, 240] loss: 0.101 [61, 300] loss: 0.098 [61, 360] loss: 0.105 Epoch: 61 -> Loss: 0.114300295711 Epoch: 61 -> Test Accuracy: 87.17 [62, 60] loss: 0.097 [62, 120] loss: 0.103 [62, 180] loss: 0.102 [62, 240] loss: 0.097 [62, 300] loss: 0.101 [62, 360] loss: 0.100 Epoch: 62 -> Loss: 0.0899142846465 Epoch: 62 -> Test Accuracy: 86.98 [63, 60] loss: 0.102 [63, 120] loss: 0.108 [63, 180] loss: 0.094 [63, 240] loss: 0.104 [63, 300] loss: 0.098 [63, 360] loss: 0.103 Epoch: 63 -> Loss: 0.134980231524 Epoch: 63 -> Test Accuracy: 87.08 [64, 60] loss: 0.103 [64, 120] loss: 0.099 [64, 180] loss: 0.097 [64, 240] loss: 0.097 [64, 300] loss: 0.097 [64, 360] loss: 0.096 Epoch: 64 -> Loss: 0.0731904357672 Epoch: 64 -> Test Accuracy: 87.04 [65, 60] loss: 0.093 [65, 120] loss: 0.104 [65, 180] loss: 0.094 [65, 240] loss: 0.098 [65, 300] loss: 0.091 [65, 360] loss: 0.101 Epoch: 65 -> Loss: 0.134622067213 Epoch: 65 -> Test Accuracy: 87.15 [66, 60] loss: 0.095 [66, 120] loss: 0.091 [66, 180] loss: 0.094 [66, 240] loss: 0.100 [66, 300] loss: 0.092 [66, 360] loss: 0.099 Epoch: 66 -> Loss: 0.0725847557187 Epoch: 66 -> Test Accuracy: 87.05 [67, 60] loss: 0.098 [67, 120] loss: 0.101 [67, 180] loss: 0.096 [67, 240] loss: 0.093 [67, 300] loss: 0.095 [67, 360] loss: 0.095 Epoch: 67 -> Loss: 0.119690179825 Epoch: 67 -> Test Accuracy: 87.04 [68, 60] loss: 0.095 [68, 120] loss: 0.090 [68, 180] loss: 0.097 [68, 240] loss: 0.096 [68, 300] loss: 0.097 [68, 360] loss: 0.098 Epoch: 68 -> Loss: 0.229573771358 Epoch: 68 -> Test Accuracy: 87.05 [69, 60] loss: 0.091 [69, 120] loss: 0.102 [69, 180] loss: 0.096 [69, 240] loss: 0.090 [69, 300] loss: 0.089 [69, 360] loss: 0.099 Epoch: 69 -> Loss: 0.0802088752389 Epoch: 69 -> Test Accuracy: 87.03 [70, 60] loss: 0.099 [70, 120] loss: 0.098 [70, 180] loss: 0.087 [70, 240] loss: 0.087 [70, 300] loss: 0.094 [70, 360] loss: 0.098 Epoch: 70 -> Loss: 0.0723464339972 Epoch: 70 -> Test Accuracy: 87.06 [71, 60] loss: 0.090 [71, 120] loss: 0.098 [71, 180] loss: 0.091 [71, 240] loss: 0.097 [71, 300] loss: 0.098 [71, 360] loss: 0.092 Epoch: 71 -> Loss: 0.0418816134334 Epoch: 71 -> Test Accuracy: 87.12 [72, 60] loss: 0.091 [72, 120] loss: 0.097 [72, 180] loss: 0.088 [72, 240] loss: 0.088 [72, 300] loss: 0.083 [72, 360] loss: 0.096 Epoch: 72 -> Loss: 0.161606714129 Epoch: 72 -> Test Accuracy: 86.95 [73, 60] loss: 0.092 [73, 120] loss: 0.090 [73, 180] loss: 0.093 [73, 240] loss: 0.098 [73, 300] loss: 0.094 [73, 360] loss: 0.088 Epoch: 73 -> Loss: 0.0995782464743 Epoch: 73 -> Test Accuracy: 87.03 [74, 60] loss: 0.088 [74, 120] loss: 0.096 [74, 180] loss: 0.094 [74, 240] loss: 0.091 [74, 300] loss: 0.092 [74, 360] loss: 0.092 Epoch: 74 -> Loss: 0.104454539716 Epoch: 74 -> Test Accuracy: 87.18 [75, 60] loss: 0.096 [75, 120] loss: 0.094 [75, 180] loss: 0.087 [75, 240] loss: 0.084 [75, 300] loss: 0.094 [75, 360] loss: 0.090 Epoch: 75 -> Loss: 0.103685036302 Epoch: 75 -> Test Accuracy: 87.02 [76, 60] loss: 0.088 [76, 120] loss: 0.083 [76, 180] loss: 0.091 [76, 240] loss: 0.090 [76, 300] loss: 0.089 [76, 360] loss: 0.086 Epoch: 76 -> Loss: 0.171830669045 Epoch: 76 -> Test Accuracy: 87.19 [77, 60] loss: 0.087 [77, 120] loss: 0.089 [77, 180] loss: 0.092 [77, 240] loss: 0.089 [77, 300] loss: 0.092 [77, 360] loss: 0.090 Epoch: 77 -> Loss: 0.0278023779392 Epoch: 77 -> Test Accuracy: 87.04 [78, 60] loss: 0.090 [78, 120] loss: 0.084 [78, 180] loss: 0.085 [78, 240] loss: 0.091 [78, 300] loss: 0.101 [78, 360] loss: 0.090 Epoch: 78 -> Loss: 0.163371950388 Epoch: 78 -> Test Accuracy: 87.17 [79, 60] loss: 0.089 [79, 120] loss: 0.095 [79, 180] loss: 0.090 [79, 240] loss: 0.086 [79, 300] loss: 0.093 [79, 360] loss: 0.089 Epoch: 79 -> Loss: 0.116952084005 Epoch: 79 -> Test Accuracy: 87.15 [80, 60] loss: 0.089 [80, 120] loss: 0.091 [80, 180] loss: 0.087 [80, 240] loss: 0.088 [80, 300] loss: 0.092 [80, 360] loss: 0.088 Epoch: 80 -> Loss: 0.177693337202 Epoch: 80 -> Test Accuracy: 87.0 [81, 60] loss: 0.085 [81, 120] loss: 0.092 [81, 180] loss: 0.085 [81, 240] loss: 0.081 [81, 300] loss: 0.092 [81, 360] loss: 0.083 Epoch: 81 -> Loss: 0.0867458954453 Epoch: 81 -> Test Accuracy: 86.99 [82, 60] loss: 0.090 [82, 120] loss: 0.097 [82, 180] loss: 0.082 [82, 240] loss: 0.090 [82, 300] loss: 0.089 [82, 360] loss: 0.088 Epoch: 82 -> Loss: 0.0585666783154 Epoch: 82 -> Test Accuracy: 87.0 [83, 60] loss: 0.084 [83, 120] loss: 0.083 [83, 180] loss: 0.093 [83, 240] loss: 0.091 [83, 300] loss: 0.086 [83, 360] loss: 0.083 Epoch: 83 -> Loss: 0.0910839065909 Epoch: 83 -> Test Accuracy: 87.12 [84, 60] loss: 0.085 [84, 120] loss: 0.082 [84, 180] loss: 0.083 [84, 240] loss: 0.085 [84, 300] loss: 0.082 [84, 360] loss: 0.075 Epoch: 84 -> Loss: 0.049478083849 Epoch: 84 -> Test Accuracy: 86.98 [85, 60] loss: 0.087 [85, 120] loss: 0.084 [85, 180] loss: 0.079 [85, 240] loss: 0.083 [85, 300] loss: 0.079 [85, 360] loss: 0.089 Epoch: 85 -> Loss: 0.0535939820111 Epoch: 85 -> Test Accuracy: 86.85 [86, 60] loss: 0.085 [86, 120] loss: 0.081 [86, 180] loss: 0.081 [86, 240] loss: 0.088 [86, 300] loss: 0.080 [86, 360] loss: 0.078 Epoch: 86 -> Loss: 0.0802417322993 Epoch: 86 -> Test Accuracy: 86.97 [87, 60] loss: 0.081 [87, 120] loss: 0.085 [87, 180] loss: 0.081 [87, 240] loss: 0.083 [87, 300] loss: 0.082 [87, 360] loss: 0.080 Epoch: 87 -> Loss: 0.237124204636 Epoch: 87 -> Test Accuracy: 86.95 [88, 60] loss: 0.082 [88, 120] loss: 0.081 [88, 180] loss: 0.082 [88, 240] loss: 0.081 [88, 300] loss: 0.080 [88, 360] loss: 0.085 Epoch: 88 -> Loss: 0.0808196440339 Epoch: 88 -> Test Accuracy: 86.99 [89, 60] loss: 0.082 [89, 120] loss: 0.080 [89, 180] loss: 0.088 [89, 240] loss: 0.081 [89, 300] loss: 0.083 [89, 360] loss: 0.080 Epoch: 89 -> Loss: 0.0974667519331 Epoch: 89 -> Test Accuracy: 86.9 [90, 60] loss: 0.084 [90, 120] loss: 0.080 [90, 180] loss: 0.080 [90, 240] loss: 0.084 [90, 300] loss: 0.084 [90, 360] loss: 0.081 Epoch: 90 -> Loss: 0.139950841665 Epoch: 90 -> Test Accuracy: 87.1 [91, 60] loss: 0.077 [91, 120] loss: 0.076 [91, 180] loss: 0.081 [91, 240] loss: 0.083 [91, 300] loss: 0.086 [91, 360] loss: 0.083 Epoch: 91 -> Loss: 0.102676652372 Epoch: 91 -> Test Accuracy: 86.91 [92, 60] loss: 0.084 [92, 120] loss: 0.075 [92, 180] loss: 0.080 [92, 240] loss: 0.081 [92, 300] loss: 0.085 [92, 360] loss: 0.083 Epoch: 92 -> Loss: 0.184652641416 Epoch: 92 -> Test Accuracy: 87.05 [93, 60] loss: 0.085 [93, 120] loss: 0.078 [93, 180] loss: 0.078 [93, 240] loss: 0.076 [93, 300] loss: 0.086 [93, 360] loss: 0.086 Epoch: 93 -> Loss: 0.121937416494 Epoch: 93 -> Test Accuracy: 87.01 [94, 60] loss: 0.087 [94, 120] loss: 0.077 [94, 180] loss: 0.082 [94, 240] loss: 0.084 [94, 300] loss: 0.078 [94, 360] loss: 0.086 Epoch: 94 -> Loss: 0.0756267905235 Epoch: 94 -> Test Accuracy: 86.91 [95, 60] loss: 0.081 [95, 120] loss: 0.073 [95, 180] loss: 0.075 [95, 240] loss: 0.086 [95, 300] loss: 0.083 [95, 360] loss: 0.080 Epoch: 95 -> Loss: 0.0927665904164 Epoch: 95 -> Test Accuracy: 87.07 [96, 60] loss: 0.076 [96, 120] loss: 0.075 [96, 180] loss: 0.079 [96, 240] loss: 0.077 [96, 300] loss: 0.085 [96, 360] loss: 0.084 Epoch: 96 -> Loss: 0.0530608296394 Epoch: 96 -> Test Accuracy: 86.98 [97, 60] loss: 0.082 [97, 120] loss: 0.083 [97, 180] loss: 0.086 [97, 240] loss: 0.074 [97, 300] loss: 0.079 [97, 360] loss: 0.075 Epoch: 97 -> Loss: 0.116631627083 Epoch: 97 -> Test Accuracy: 86.9 [98, 60] loss: 0.078 [98, 120] loss: 0.072 [98, 180] loss: 0.080 [98, 240] loss: 0.080 [98, 300] loss: 0.071 [98, 360] loss: 0.079 Epoch: 98 -> Loss: 0.0896067619324 Epoch: 98 -> Test Accuracy: 86.88 [99, 60] loss: 0.073 [99, 120] loss: 0.077 [99, 180] loss: 0.078 [99, 240] loss: 0.079 [99, 300] loss: 0.076 [99, 360] loss: 0.079 Epoch: 99 -> Loss: 0.0433508194983 Epoch: 99 -> Test Accuracy: 86.93 [100, 60] loss: 0.074 [100, 120] loss: 0.076 [100, 180] loss: 0.074 [100, 240] loss: 0.076 [100, 300] loss: 0.066 [100, 360] loss: 0.078 Epoch: 100 -> Loss: 0.0526957735419 Epoch: 100 -> Test Accuracy: 86.98 Finished Training [1, 60] loss: 1.605 [1, 120] loss: 0.846 [1, 180] loss: 0.789 [1, 240] loss: 0.733 [1, 300] loss: 0.681 [1, 360] loss: 0.677 Epoch: 1 -> Loss: 0.769748389721 Epoch: 1 -> Test Accuracy: 74.84 [2, 60] loss: 0.651 [2, 120] loss: 0.633 [2, 180] loss: 0.632 [2, 240] loss: 0.607 [2, 300] loss: 0.605 [2, 360] loss: 0.616 Epoch: 2 -> Loss: 0.742862284184 Epoch: 2 -> Test Accuracy: 77.06 [3, 60] loss: 0.568 [3, 120] loss: 0.549 [3, 180] loss: 0.576 [3, 240] loss: 0.564 [3, 300] loss: 0.575 [3, 360] loss: 0.577 Epoch: 3 -> Loss: 0.471496880054 Epoch: 3 -> Test Accuracy: 79.05 [4, 60] loss: 0.523 [4, 120] loss: 0.529 [4, 180] loss: 0.523 [4, 240] loss: 0.545 [4, 300] loss: 0.549 [4, 360] loss: 0.539 Epoch: 4 -> Loss: 0.706794142723 Epoch: 4 -> Test Accuracy: 79.15 [5, 60] loss: 0.503 [5, 120] loss: 0.496 [5, 180] loss: 0.533 [5, 240] loss: 0.522 [5, 300] loss: 0.519 [5, 360] loss: 0.508 Epoch: 5 -> Loss: 0.490466743708 Epoch: 5 -> Test Accuracy: 79.89 [6, 60] loss: 0.473 [6, 120] loss: 0.499 [6, 180] loss: 0.504 [6, 240] loss: 0.515 [6, 300] loss: 0.498 [6, 360] loss: 0.493 Epoch: 6 -> Loss: 0.545969605446 Epoch: 6 -> Test Accuracy: 79.9 [7, 60] loss: 0.492 [7, 120] loss: 0.469 [7, 180] loss: 0.488 [7, 240] loss: 0.483 [7, 300] loss: 0.517 [7, 360] loss: 0.486 Epoch: 7 -> Loss: 0.551766335964 Epoch: 7 -> Test Accuracy: 80.26 [8, 60] loss: 0.477 [8, 120] loss: 0.478 [8, 180] loss: 0.496 [8, 240] loss: 0.461 [8, 300] loss: 0.504 [8, 360] loss: 0.487 Epoch: 8 -> Loss: 0.427379071712 Epoch: 8 -> Test Accuracy: 80.36 [9, 60] loss: 0.472 [9, 120] loss: 0.462 [9, 180] loss: 0.465 [9, 240] loss: 0.477 [9, 300] loss: 0.477 [9, 360] loss: 0.478 Epoch: 9 -> Loss: 0.429799467325 Epoch: 9 -> Test Accuracy: 79.96 [10, 60] loss: 0.454 [10, 120] loss: 0.465 [10, 180] loss: 0.463 [10, 240] loss: 0.471 [10, 300] loss: 0.463 [10, 360] loss: 0.464 Epoch: 10 -> Loss: 0.401413530111 Epoch: 10 -> Test Accuracy: 80.69 [11, 60] loss: 0.444 [11, 120] loss: 0.448 [11, 180] loss: 0.444 [11, 240] loss: 0.481 [11, 300] loss: 0.482 [11, 360] loss: 0.477 Epoch: 11 -> Loss: 0.395341336727 Epoch: 11 -> Test Accuracy: 80.08 [12, 60] loss: 0.434 [12, 120] loss: 0.448 [12, 180] loss: 0.454 [12, 240] loss: 0.459 [12, 300] loss: 0.473 [12, 360] loss: 0.464 Epoch: 12 -> Loss: 0.510496497154 Epoch: 12 -> Test Accuracy: 80.42 [13, 60] loss: 0.430 [13, 120] loss: 0.435 [13, 180] loss: 0.457 [13, 240] loss: 0.459 [13, 300] loss: 0.465 [13, 360] loss: 0.451 Epoch: 13 -> Loss: 0.47343057394 Epoch: 13 -> Test Accuracy: 80.07 [14, 60] loss: 0.434 [14, 120] loss: 0.442 [14, 180] loss: 0.437 [14, 240] loss: 0.448 [14, 300] loss: 0.445 [14, 360] loss: 0.454 Epoch: 14 -> Loss: 0.414171218872 Epoch: 14 -> Test Accuracy: 80.18 [15, 60] loss: 0.432 [15, 120] loss: 0.448 [15, 180] loss: 0.452 [15, 240] loss: 0.448 [15, 300] loss: 0.451 [15, 360] loss: 0.453 Epoch: 15 -> Loss: 0.561669826508 Epoch: 15 -> Test Accuracy: 80.31 [16, 60] loss: 0.438 [16, 120] loss: 0.429 [16, 180] loss: 0.430 [16, 240] loss: 0.450 [16, 300] loss: 0.457 [16, 360] loss: 0.454 Epoch: 16 -> Loss: 0.567278921604 Epoch: 16 -> Test Accuracy: 80.55 [17, 60] loss: 0.429 [17, 120] loss: 0.418 [17, 180] loss: 0.448 [17, 240] loss: 0.437 [17, 300] loss: 0.462 [17, 360] loss: 0.465 Epoch: 17 -> Loss: 0.570365846157 Epoch: 17 -> Test Accuracy: 80.12 [18, 60] loss: 0.417 [18, 120] loss: 0.419 [18, 180] loss: 0.441 [18, 240] loss: 0.447 [18, 300] loss: 0.458 [18, 360] loss: 0.457 Epoch: 18 -> Loss: 0.421803236008 Epoch: 18 -> Test Accuracy: 80.08 [19, 60] loss: 0.415 [19, 120] loss: 0.421 [19, 180] loss: 0.433 [19, 240] loss: 0.439 [19, 300] loss: 0.441 [19, 360] loss: 0.444 Epoch: 19 -> Loss: 0.302561104298 Epoch: 19 -> Test Accuracy: 80.31 [20, 60] loss: 0.421 [20, 120] loss: 0.416 [20, 180] loss: 0.429 [20, 240] loss: 0.437 [20, 300] loss: 0.443 [20, 360] loss: 0.471 Epoch: 20 -> Loss: 0.446769177914 Epoch: 20 -> Test Accuracy: 80.8 [21, 60] loss: 0.377 [21, 120] loss: 0.370 [21, 180] loss: 0.358 [21, 240] loss: 0.392 [21, 300] loss: 0.357 [21, 360] loss: 0.363 Epoch: 21 -> Loss: 0.316772520542 Epoch: 21 -> Test Accuracy: 82.42 [22, 60] loss: 0.336 [22, 120] loss: 0.334 [22, 180] loss: 0.342 [22, 240] loss: 0.346 [22, 300] loss: 0.342 [22, 360] loss: 0.336 Epoch: 22 -> Loss: 0.357571542263 Epoch: 22 -> Test Accuracy: 82.4 [23, 60] loss: 0.322 [23, 120] loss: 0.326 [23, 180] loss: 0.321 [23, 240] loss: 0.319 [23, 300] loss: 0.322 [23, 360] loss: 0.328 Epoch: 23 -> Loss: 0.370080560446 Epoch: 23 -> Test Accuracy: 82.57 [24, 60] loss: 0.310 [24, 120] loss: 0.316 [24, 180] loss: 0.314 [24, 240] loss: 0.321 [24, 300] loss: 0.315 [24, 360] loss: 0.321 Epoch: 24 -> Loss: 0.189372211695 Epoch: 24 -> Test Accuracy: 82.79 [25, 60] loss: 0.301 [25, 120] loss: 0.307 [25, 180] loss: 0.297 [25, 240] loss: 0.307 [25, 300] loss: 0.317 [25, 360] loss: 0.303 Epoch: 25 -> Loss: 0.396466910839 Epoch: 25 -> Test Accuracy: 82.75 [26, 60] loss: 0.292 [26, 120] loss: 0.299 [26, 180] loss: 0.301 [26, 240] loss: 0.315 [26, 300] loss: 0.304 [26, 360] loss: 0.302 Epoch: 26 -> Loss: 0.238028690219 Epoch: 26 -> Test Accuracy: 82.81 [27, 60] loss: 0.293 [27, 120] loss: 0.296 [27, 180] loss: 0.285 [27, 240] loss: 0.299 [27, 300] loss: 0.295 [27, 360] loss: 0.288 Epoch: 27 -> Loss: 0.278031021357 Epoch: 27 -> Test Accuracy: 82.71 [28, 60] loss: 0.291 [28, 120] loss: 0.287 [28, 180] loss: 0.283 [28, 240] loss: 0.302 [28, 300] loss: 0.269 [28, 360] loss: 0.292 Epoch: 28 -> Loss: 0.435865014791 Epoch: 28 -> Test Accuracy: 82.66 [29, 60] loss: 0.291 [29, 120] loss: 0.280 [29, 180] loss: 0.283 [29, 240] loss: 0.295 [29, 300] loss: 0.295 [29, 360] loss: 0.291 Epoch: 29 -> Loss: 0.275308996439 Epoch: 29 -> Test Accuracy: 82.79 [30, 60] loss: 0.280 [30, 120] loss: 0.270 [30, 180] loss: 0.277 [30, 240] loss: 0.294 [30, 300] loss: 0.288 [30, 360] loss: 0.301 Epoch: 30 -> Loss: 0.280865430832 Epoch: 30 -> Test Accuracy: 82.21 [31, 60] loss: 0.277 [31, 120] loss: 0.284 [31, 180] loss: 0.283 [31, 240] loss: 0.291 [31, 300] loss: 0.271 [31, 360] loss: 0.297 Epoch: 31 -> Loss: 0.391873121262 Epoch: 31 -> Test Accuracy: 82.69 [32, 60] loss: 0.269 [32, 120] loss: 0.289 [32, 180] loss: 0.264 [32, 240] loss: 0.282 [32, 300] loss: 0.289 [32, 360] loss: 0.291 Epoch: 32 -> Loss: 0.343095004559 Epoch: 32 -> Test Accuracy: 82.28 [33, 60] loss: 0.278 [33, 120] loss: 0.273 [33, 180] loss: 0.273 [33, 240] loss: 0.281 [33, 300] loss: 0.285 [33, 360] loss: 0.280 Epoch: 33 -> Loss: 0.279102951288 Epoch: 33 -> Test Accuracy: 81.93 [34, 60] loss: 0.273 [34, 120] loss: 0.266 [34, 180] loss: 0.292 [34, 240] loss: 0.283 [34, 300] loss: 0.285 [34, 360] loss: 0.272 Epoch: 34 -> Loss: 0.251658469439 Epoch: 34 -> Test Accuracy: 82.35 [35, 60] loss: 0.260 [35, 120] loss: 0.278 [35, 180] loss: 0.274 [35, 240] loss: 0.278 [35, 300] loss: 0.288 [35, 360] loss: 0.299 Epoch: 35 -> Loss: 0.244260117412 Epoch: 35 -> Test Accuracy: 82.33 [36, 60] loss: 0.267 [36, 120] loss: 0.268 [36, 180] loss: 0.270 [36, 240] loss: 0.296 [36, 300] loss: 0.266 [36, 360] loss: 0.285 Epoch: 36 -> Loss: 0.333917081356 Epoch: 36 -> Test Accuracy: 81.92 [37, 60] loss: 0.269 [37, 120] loss: 0.261 [37, 180] loss: 0.284 [37, 240] loss: 0.262 [37, 300] loss: 0.286 [37, 360] loss: 0.282 Epoch: 37 -> Loss: 0.405453115702 Epoch: 37 -> Test Accuracy: 82.28 [38, 60] loss: 0.252 [38, 120] loss: 0.272 [38, 180] loss: 0.263 [38, 240] loss: 0.270 [38, 300] loss: 0.275 [38, 360] loss: 0.273 Epoch: 38 -> Loss: 0.405525028706 Epoch: 38 -> Test Accuracy: 81.91 [39, 60] loss: 0.267 [39, 120] loss: 0.252 [39, 180] loss: 0.268 [39, 240] loss: 0.268 [39, 300] loss: 0.290 [39, 360] loss: 0.297 Epoch: 39 -> Loss: 0.246826142073 Epoch: 39 -> Test Accuracy: 82.42 [40, 60] loss: 0.269 [40, 120] loss: 0.268 [40, 180] loss: 0.261 [40, 240] loss: 0.260 [40, 300] loss: 0.257 [40, 360] loss: 0.284 Epoch: 40 -> Loss: 0.45331415534 Epoch: 40 -> Test Accuracy: 82.42 [41, 60] loss: 0.236 [41, 120] loss: 0.252 [41, 180] loss: 0.250 [41, 240] loss: 0.238 [41, 300] loss: 0.241 [41, 360] loss: 0.232 Epoch: 41 -> Loss: 0.207646131516 Epoch: 41 -> Test Accuracy: 82.86 [42, 60] loss: 0.222 [42, 120] loss: 0.233 [42, 180] loss: 0.216 [42, 240] loss: 0.213 [42, 300] loss: 0.231 [42, 360] loss: 0.223 Epoch: 42 -> Loss: 0.273645073175 Epoch: 42 -> Test Accuracy: 83.3 [43, 60] loss: 0.206 [43, 120] loss: 0.220 [43, 180] loss: 0.208 [43, 240] loss: 0.215 [43, 300] loss: 0.220 [43, 360] loss: 0.209 Epoch: 43 -> Loss: 0.0931783020496 Epoch: 43 -> Test Accuracy: 83.09 [44, 60] loss: 0.207 [44, 120] loss: 0.201 [44, 180] loss: 0.207 [44, 240] loss: 0.214 [44, 300] loss: 0.209 [44, 360] loss: 0.196 Epoch: 44 -> Loss: 0.167782276869 Epoch: 44 -> Test Accuracy: 83.05 [45, 60] loss: 0.202 [45, 120] loss: 0.194 [45, 180] loss: 0.199 [45, 240] loss: 0.191 [45, 300] loss: 0.199 [45, 360] loss: 0.198 Epoch: 45 -> Loss: 0.162827700377 Epoch: 45 -> Test Accuracy: 83.47 [46, 60] loss: 0.189 [46, 120] loss: 0.192 [46, 180] loss: 0.190 [46, 240] loss: 0.191 [46, 300] loss: 0.195 [46, 360] loss: 0.198 Epoch: 46 -> Loss: 0.235953241587 Epoch: 46 -> Test Accuracy: 83.36 [47, 60] loss: 0.194 [47, 120] loss: 0.194 [47, 180] loss: 0.188 [47, 240] loss: 0.182 [47, 300] loss: 0.188 [47, 360] loss: 0.194 Epoch: 47 -> Loss: 0.223013162613 Epoch: 47 -> Test Accuracy: 83.47 [48, 60] loss: 0.180 [48, 120] loss: 0.190 [48, 180] loss: 0.181 [48, 240] loss: 0.187 [48, 300] loss: 0.199 [48, 360] loss: 0.190 Epoch: 48 -> Loss: 0.139108017087 Epoch: 48 -> Test Accuracy: 83.31 [49, 60] loss: 0.190 [49, 120] loss: 0.187 [49, 180] loss: 0.179 [49, 240] loss: 0.183 [49, 300] loss: 0.194 [49, 360] loss: 0.188 Epoch: 49 -> Loss: 0.14915433526 Epoch: 49 -> Test Accuracy: 83.36 [50, 60] loss: 0.183 [50, 120] loss: 0.182 [50, 180] loss: 0.194 [50, 240] loss: 0.190 [50, 300] loss: 0.186 [50, 360] loss: 0.180 Epoch: 50 -> Loss: 0.106967367232 Epoch: 50 -> Test Accuracy: 83.36 [51, 60] loss: 0.186 [51, 120] loss: 0.175 [51, 180] loss: 0.182 [51, 240] loss: 0.186 [51, 300] loss: 0.181 [51, 360] loss: 0.183 Epoch: 51 -> Loss: 0.187744662166 Epoch: 51 -> Test Accuracy: 83.28 [52, 60] loss: 0.177 [52, 120] loss: 0.176 [52, 180] loss: 0.184 [52, 240] loss: 0.188 [52, 300] loss: 0.186 [52, 360] loss: 0.174 Epoch: 52 -> Loss: 0.201156467199 Epoch: 52 -> Test Accuracy: 83.39 [53, 60] loss: 0.173 [53, 120] loss: 0.184 [53, 180] loss: 0.180 [53, 240] loss: 0.188 [53, 300] loss: 0.188 [53, 360] loss: 0.172 Epoch: 53 -> Loss: 0.0869134441018 Epoch: 53 -> Test Accuracy: 83.35 [54, 60] loss: 0.180 [54, 120] loss: 0.173 [54, 180] loss: 0.182 [54, 240] loss: 0.170 [54, 300] loss: 0.177 [54, 360] loss: 0.168 Epoch: 54 -> Loss: 0.182309672236 Epoch: 54 -> Test Accuracy: 83.37 [55, 60] loss: 0.175 [55, 120] loss: 0.175 [55, 180] loss: 0.178 [55, 240] loss: 0.197 [55, 300] loss: 0.170 [55, 360] loss: 0.174 Epoch: 55 -> Loss: 0.215438812971 Epoch: 55 -> Test Accuracy: 83.33 [56, 60] loss: 0.178 [56, 120] loss: 0.180 [56, 180] loss: 0.178 [56, 240] loss: 0.172 [56, 300] loss: 0.181 [56, 360] loss: 0.165 Epoch: 56 -> Loss: 0.229026034474 Epoch: 56 -> Test Accuracy: 83.44 [57, 60] loss: 0.173 [57, 120] loss: 0.175 [57, 180] loss: 0.180 [57, 240] loss: 0.171 [57, 300] loss: 0.178 [57, 360] loss: 0.176 Epoch: 57 -> Loss: 0.236210376024 Epoch: 57 -> Test Accuracy: 83.46 [58, 60] loss: 0.173 [58, 120] loss: 0.177 [58, 180] loss: 0.173 [58, 240] loss: 0.177 [58, 300] loss: 0.175 [58, 360] loss: 0.181 Epoch: 58 -> Loss: 0.195679098368 Epoch: 58 -> Test Accuracy: 83.43 [59, 60] loss: 0.171 [59, 120] loss: 0.172 [59, 180] loss: 0.176 [59, 240] loss: 0.173 [59, 300] loss: 0.169 [59, 360] loss: 0.167 Epoch: 59 -> Loss: 0.11266014725 Epoch: 59 -> Test Accuracy: 83.51 [60, 60] loss: 0.170 [60, 120] loss: 0.180 [60, 180] loss: 0.162 [60, 240] loss: 0.170 [60, 300] loss: 0.166 [60, 360] loss: 0.177 Epoch: 60 -> Loss: 0.191300824285 Epoch: 60 -> Test Accuracy: 83.55 [61, 60] loss: 0.162 [61, 120] loss: 0.162 [61, 180] loss: 0.174 [61, 240] loss: 0.174 [61, 300] loss: 0.175 [61, 360] loss: 0.171 Epoch: 61 -> Loss: 0.285082429647 Epoch: 61 -> Test Accuracy: 83.53 [62, 60] loss: 0.166 [62, 120] loss: 0.179 [62, 180] loss: 0.171 [62, 240] loss: 0.156 [62, 300] loss: 0.167 [62, 360] loss: 0.165 Epoch: 62 -> Loss: 0.17471639812 Epoch: 62 -> Test Accuracy: 83.68 [63, 60] loss: 0.166 [63, 120] loss: 0.169 [63, 180] loss: 0.168 [63, 240] loss: 0.164 [63, 300] loss: 0.163 [63, 360] loss: 0.159 Epoch: 63 -> Loss: 0.27798050642 Epoch: 63 -> Test Accuracy: 83.62 [64, 60] loss: 0.172 [64, 120] loss: 0.171 [64, 180] loss: 0.169 [64, 240] loss: 0.171 [64, 300] loss: 0.173 [64, 360] loss: 0.166 Epoch: 64 -> Loss: 0.182154223323 Epoch: 64 -> Test Accuracy: 83.73 [65, 60] loss: 0.178 [65, 120] loss: 0.171 [65, 180] loss: 0.164 [65, 240] loss: 0.161 [65, 300] loss: 0.167 [65, 360] loss: 0.165 Epoch: 65 -> Loss: 0.278418779373 Epoch: 65 -> Test Accuracy: 83.6 [66, 60] loss: 0.165 [66, 120] loss: 0.161 [66, 180] loss: 0.160 [66, 240] loss: 0.173 [66, 300] loss: 0.170 [66, 360] loss: 0.158 Epoch: 66 -> Loss: 0.153344780207 Epoch: 66 -> Test Accuracy: 83.56 [67, 60] loss: 0.157 [67, 120] loss: 0.159 [67, 180] loss: 0.173 [67, 240] loss: 0.160 [67, 300] loss: 0.176 [67, 360] loss: 0.171 Epoch: 67 -> Loss: 0.167253404856 Epoch: 67 -> Test Accuracy: 83.56 [68, 60] loss: 0.171 [68, 120] loss: 0.164 [68, 180] loss: 0.157 [68, 240] loss: 0.164 [68, 300] loss: 0.161 [68, 360] loss: 0.170 Epoch: 68 -> Loss: 0.137328147888 Epoch: 68 -> Test Accuracy: 83.62 [69, 60] loss: 0.156 [69, 120] loss: 0.153 [69, 180] loss: 0.172 [69, 240] loss: 0.165 [69, 300] loss: 0.155 [69, 360] loss: 0.162 Epoch: 69 -> Loss: 0.244012355804 Epoch: 69 -> Test Accuracy: 83.62 [70, 60] loss: 0.168 [70, 120] loss: 0.172 [70, 180] loss: 0.155 [70, 240] loss: 0.156 [70, 300] loss: 0.169 [70, 360] loss: 0.165 Epoch: 70 -> Loss: 0.161557644606 Epoch: 70 -> Test Accuracy: 83.72 [71, 60] loss: 0.166 [71, 120] loss: 0.160 [71, 180] loss: 0.154 [71, 240] loss: 0.169 [71, 300] loss: 0.159 [71, 360] loss: 0.169 Epoch: 71 -> Loss: 0.236004680395 Epoch: 71 -> Test Accuracy: 83.72 [72, 60] loss: 0.166 [72, 120] loss: 0.156 [72, 180] loss: 0.164 [72, 240] loss: 0.157 [72, 300] loss: 0.164 [72, 360] loss: 0.155 Epoch: 72 -> Loss: 0.163643166423 Epoch: 72 -> Test Accuracy: 83.76 [73, 60] loss: 0.152 [73, 120] loss: 0.157 [73, 180] loss: 0.155 [73, 240] loss: 0.161 [73, 300] loss: 0.160 [73, 360] loss: 0.156 Epoch: 73 -> Loss: 0.225978657603 Epoch: 73 -> Test Accuracy: 83.64 [74, 60] loss: 0.158 [74, 120] loss: 0.167 [74, 180] loss: 0.163 [74, 240] loss: 0.155 [74, 300] loss: 0.159 [74, 360] loss: 0.151 Epoch: 74 -> Loss: 0.10319314152 Epoch: 74 -> Test Accuracy: 83.79 [75, 60] loss: 0.163 [75, 120] loss: 0.158 [75, 180] loss: 0.159 [75, 240] loss: 0.148 [75, 300] loss: 0.153 [75, 360] loss: 0.159 Epoch: 75 -> Loss: 0.128170013428 Epoch: 75 -> Test Accuracy: 83.71 [76, 60] loss: 0.154 [76, 120] loss: 0.149 [76, 180] loss: 0.151 [76, 240] loss: 0.157 [76, 300] loss: 0.155 [76, 360] loss: 0.157 Epoch: 76 -> Loss: 0.192204624414 Epoch: 76 -> Test Accuracy: 83.75 [77, 60] loss: 0.168 [77, 120] loss: 0.157 [77, 180] loss: 0.156 [77, 240] loss: 0.151 [77, 300] loss: 0.160 [77, 360] loss: 0.163 Epoch: 77 -> Loss: 0.104353502393 Epoch: 77 -> Test Accuracy: 83.66 [78, 60] loss: 0.154 [78, 120] loss: 0.154 [78, 180] loss: 0.160 [78, 240] loss: 0.164 [78, 300] loss: 0.156 [78, 360] loss: 0.156 Epoch: 78 -> Loss: 0.0883823335171 Epoch: 78 -> Test Accuracy: 83.64 [79, 60] loss: 0.154 [79, 120] loss: 0.155 [79, 180] loss: 0.155 [79, 240] loss: 0.144 [79, 300] loss: 0.158 [79, 360] loss: 0.156 Epoch: 79 -> Loss: 0.155757188797 Epoch: 79 -> Test Accuracy: 83.64 [80, 60] loss: 0.149 [80, 120] loss: 0.151 [80, 180] loss: 0.151 [80, 240] loss: 0.163 [80, 300] loss: 0.151 [80, 360] loss: 0.151 Epoch: 80 -> Loss: 0.0908910185099 Epoch: 80 -> Test Accuracy: 83.64 [81, 60] loss: 0.158 [81, 120] loss: 0.152 [81, 180] loss: 0.155 [81, 240] loss: 0.154 [81, 300] loss: 0.149 [81, 360] loss: 0.154 Epoch: 81 -> Loss: 0.234690546989 Epoch: 81 -> Test Accuracy: 83.57 [82, 60] loss: 0.155 [82, 120] loss: 0.152 [82, 180] loss: 0.151 [82, 240] loss: 0.151 [82, 300] loss: 0.149 [82, 360] loss: 0.147 Epoch: 82 -> Loss: 0.235760167241 Epoch: 82 -> Test Accuracy: 83.59 [83, 60] loss: 0.153 [83, 120] loss: 0.151 [83, 180] loss: 0.155 [83, 240] loss: 0.156 [83, 300] loss: 0.152 [83, 360] loss: 0.149 Epoch: 83 -> Loss: 0.20064611733 Epoch: 83 -> Test Accuracy: 83.6 [84, 60] loss: 0.146 [84, 120] loss: 0.147 [84, 180] loss: 0.145 [84, 240] loss: 0.154 [84, 300] loss: 0.146 [84, 360] loss: 0.154 Epoch: 84 -> Loss: 0.202218964696 Epoch: 84 -> Test Accuracy: 83.55 [85, 60] loss: 0.151 [85, 120] loss: 0.155 [85, 180] loss: 0.156 [85, 240] loss: 0.141 [85, 300] loss: 0.150 [85, 360] loss: 0.149 Epoch: 85 -> Loss: 0.15076392889 Epoch: 85 -> Test Accuracy: 83.65 [86, 60] loss: 0.151 [86, 120] loss: 0.152 [86, 180] loss: 0.146 [86, 240] loss: 0.149 [86, 300] loss: 0.146 [86, 360] loss: 0.151 Epoch: 86 -> Loss: 0.150267452002 Epoch: 86 -> Test Accuracy: 83.52 [87, 60] loss: 0.145 [87, 120] loss: 0.149 [87, 180] loss: 0.149 [87, 240] loss: 0.139 [87, 300] loss: 0.147 [87, 360] loss: 0.150 Epoch: 87 -> Loss: 0.21535487473 Epoch: 87 -> Test Accuracy: 83.51 [88, 60] loss: 0.142 [88, 120] loss: 0.142 [88, 180] loss: 0.147 [88, 240] loss: 0.149 [88, 300] loss: 0.152 [88, 360] loss: 0.149 Epoch: 88 -> Loss: 0.199831798673 Epoch: 88 -> Test Accuracy: 83.53 [89, 60] loss: 0.141 [89, 120] loss: 0.142 [89, 180] loss: 0.140 [89, 240] loss: 0.144 [89, 300] loss: 0.150 [89, 360] loss: 0.154 Epoch: 89 -> Loss: 0.176730006933 Epoch: 89 -> Test Accuracy: 83.55 [90, 60] loss: 0.140 [90, 120] loss: 0.139 [90, 180] loss: 0.148 [90, 240] loss: 0.148 [90, 300] loss: 0.162 [90, 360] loss: 0.153 Epoch: 90 -> Loss: 0.120919801295 Epoch: 90 -> Test Accuracy: 83.51 [91, 60] loss: 0.146 [91, 120] loss: 0.149 [91, 180] loss: 0.143 [91, 240] loss: 0.147 [91, 300] loss: 0.150 [91, 360] loss: 0.142 Epoch: 91 -> Loss: 0.200098872185 Epoch: 91 -> Test Accuracy: 83.45 [92, 60] loss: 0.141 [92, 120] loss: 0.148 [92, 180] loss: 0.150 [92, 240] loss: 0.140 [92, 300] loss: 0.150 [92, 360] loss: 0.152 Epoch: 92 -> Loss: 0.272257626057 Epoch: 92 -> Test Accuracy: 83.67 [93, 60] loss: 0.134 [93, 120] loss: 0.146 [93, 180] loss: 0.135 [93, 240] loss: 0.156 [93, 300] loss: 0.136 [93, 360] loss: 0.141 Epoch: 93 -> Loss: 0.132149830461 Epoch: 93 -> Test Accuracy: 83.68 [94, 60] loss: 0.146 [94, 120] loss: 0.148 [94, 180] loss: 0.139 [94, 240] loss: 0.146 [94, 300] loss: 0.149 [94, 360] loss: 0.145 Epoch: 94 -> Loss: 0.156441152096 Epoch: 94 -> Test Accuracy: 83.57 [95, 60] loss: 0.136 [95, 120] loss: 0.147 [95, 180] loss: 0.145 [95, 240] loss: 0.138 [95, 300] loss: 0.146 [95, 360] loss: 0.145 Epoch: 95 -> Loss: 0.163300901651 Epoch: 95 -> Test Accuracy: 83.71 [96, 60] loss: 0.144 [96, 120] loss: 0.138 [96, 180] loss: 0.139 [96, 240] loss: 0.147 [96, 300] loss: 0.141 [96, 360] loss: 0.150 Epoch: 96 -> Loss: 0.138925388455 Epoch: 96 -> Test Accuracy: 83.62 [97, 60] loss: 0.133 [97, 120] loss: 0.148 [97, 180] loss: 0.142 [97, 240] loss: 0.141 [97, 300] loss: 0.143 [97, 360] loss: 0.144 Epoch: 97 -> Loss: 0.102161839604 Epoch: 97 -> Test Accuracy: 83.65 [98, 60] loss: 0.140 [98, 120] loss: 0.137 [98, 180] loss: 0.129 [98, 240] loss: 0.139 [98, 300] loss: 0.139 [98, 360] loss: 0.146 Epoch: 98 -> Loss: 0.102180123329 Epoch: 98 -> Test Accuracy: 83.63 [99, 60] loss: 0.140 [99, 120] loss: 0.147 [99, 180] loss: 0.137 [99, 240] loss: 0.128 [99, 300] loss: 0.141 [99, 360] loss: 0.144 Epoch: 99 -> Loss: 0.225455522537 Epoch: 99 -> Test Accuracy: 83.59 [100, 60] loss: 0.144 [100, 120] loss: 0.142 [100, 180] loss: 0.141 [100, 240] loss: 0.136 [100, 300] loss: 0.138 [100, 360] loss: 0.134 Epoch: 100 -> Loss: 0.0870344862342 Epoch: 100 -> Test Accuracy: 83.74 Finished Training [1, 60] loss: 1.992 [1, 120] loss: 1.202 [1, 180] loss: 1.107 [1, 240] loss: 1.072 [1, 300] loss: 0.994 [1, 360] loss: 1.001 Epoch: 1 -> Loss: 0.888210117817 Epoch: 1 -> Test Accuracy: 60.67 [2, 60] loss: 0.973 [2, 120] loss: 0.949 [2, 180] loss: 0.926 [2, 240] loss: 0.932 [2, 300] loss: 0.921 [2, 360] loss: 0.890 Epoch: 2 -> Loss: 0.802006244659 Epoch: 2 -> Test Accuracy: 63.53 [3, 60] loss: 0.880 [3, 120] loss: 0.873 [3, 180] loss: 0.894 [3, 240] loss: 0.856 [3, 300] loss: 0.872 [3, 360] loss: 0.861 Epoch: 3 -> Loss: 0.733943283558 Epoch: 3 -> Test Accuracy: 65.67 [4, 60] loss: 0.836 [4, 120] loss: 0.848 [4, 180] loss: 0.838 [4, 240] loss: 0.841 [4, 300] loss: 0.839 [4, 360] loss: 0.840 Epoch: 4 -> Loss: 0.755034208298 Epoch: 4 -> Test Accuracy: 66.71 [5, 60] loss: 0.832 [5, 120] loss: 0.817 [5, 180] loss: 0.815 [5, 240] loss: 0.818 [5, 300] loss: 0.834 [5, 360] loss: 0.807 Epoch: 5 -> Loss: 0.732703328133 Epoch: 5 -> Test Accuracy: 66.19 [6, 60] loss: 0.810 [6, 120] loss: 0.808 [6, 180] loss: 0.836 [6, 240] loss: 0.827 [6, 300] loss: 0.813 [6, 360] loss: 0.792 Epoch: 6 -> Loss: 0.864481449127 Epoch: 6 -> Test Accuracy: 67.37 [7, 60] loss: 0.809 [7, 120] loss: 0.803 [7, 180] loss: 0.805 [7, 240] loss: 0.806 [7, 300] loss: 0.797 [7, 360] loss: 0.794 Epoch: 7 -> Loss: 1.01008820534 Epoch: 7 -> Test Accuracy: 67.29 [8, 60] loss: 0.791 [8, 120] loss: 0.785 [8, 180] loss: 0.792 [8, 240] loss: 0.802 [8, 300] loss: 0.803 [8, 360] loss: 0.793 Epoch: 8 -> Loss: 0.807754516602 Epoch: 8 -> Test Accuracy: 67.06 [9, 60] loss: 0.787 [9, 120] loss: 0.795 [9, 180] loss: 0.793 [9, 240] loss: 0.779 [9, 300] loss: 0.779 [9, 360] loss: 0.787 Epoch: 9 -> Loss: 0.865053653717 Epoch: 9 -> Test Accuracy: 67.92 [10, 60] loss: 0.798 [10, 120] loss: 0.774 [10, 180] loss: 0.772 [10, 240] loss: 0.778 [10, 300] loss: 0.797 [10, 360] loss: 0.789 Epoch: 10 -> Loss: 1.00264394283 Epoch: 10 -> Test Accuracy: 67.31 [11, 60] loss: 0.765 [11, 120] loss: 0.788 [11, 180] loss: 0.776 [11, 240] loss: 0.780 [11, 300] loss: 0.789 [11, 360] loss: 0.786 Epoch: 11 -> Loss: 0.86000585556 Epoch: 11 -> Test Accuracy: 67.72 [12, 60] loss: 0.747 [12, 120] loss: 0.782 [12, 180] loss: 0.759 [12, 240] loss: 0.779 [12, 300] loss: 0.804 [12, 360] loss: 0.799 Epoch: 12 -> Loss: 0.719414174557 Epoch: 12 -> Test Accuracy: 68.17 [13, 60] loss: 0.775 [13, 120] loss: 0.765 [13, 180] loss: 0.778 [13, 240] loss: 0.779 [13, 300] loss: 0.783 [13, 360] loss: 0.771 Epoch: 13 -> Loss: 0.556922793388 Epoch: 13 -> Test Accuracy: 67.7 [14, 60] loss: 0.767 [14, 120] loss: 0.781 [14, 180] loss: 0.786 [14, 240] loss: 0.771 [14, 300] loss: 0.760 [14, 360] loss: 0.791 Epoch: 14 -> Loss: 0.718975305557 Epoch: 14 -> Test Accuracy: 68.25 [15, 60] loss: 0.766 [15, 120] loss: 0.771 [15, 180] loss: 0.760 [15, 240] loss: 0.768 [15, 300] loss: 0.761 [15, 360] loss: 0.762 Epoch: 15 -> Loss: 1.00635278225 Epoch: 15 -> Test Accuracy: 68.01 [16, 60] loss: 0.752 [16, 120] loss: 0.779 [16, 180] loss: 0.773 [16, 240] loss: 0.779 [16, 300] loss: 0.761 [16, 360] loss: 0.764 Epoch: 16 -> Loss: 0.615387439728 Epoch: 16 -> Test Accuracy: 68.28 [17, 60] loss: 0.756 [17, 120] loss: 0.766 [17, 180] loss: 0.762 [17, 240] loss: 0.770 [17, 300] loss: 0.766 [17, 360] loss: 0.778 Epoch: 17 -> Loss: 0.857590973377 Epoch: 17 -> Test Accuracy: 67.68 [18, 60] loss: 0.747 [18, 120] loss: 0.748 [18, 180] loss: 0.773 [18, 240] loss: 0.784 [18, 300] loss: 0.758 [18, 360] loss: 0.759 Epoch: 18 -> Loss: 0.73320287466 Epoch: 18 -> Test Accuracy: 68.2 [19, 60] loss: 0.761 [19, 120] loss: 0.769 [19, 180] loss: 0.747 [19, 240] loss: 0.765 [19, 300] loss: 0.754 [19, 360] loss: 0.760 Epoch: 19 -> Loss: 0.831901073456 Epoch: 19 -> Test Accuracy: 68.74 [20, 60] loss: 0.757 [20, 120] loss: 0.756 [20, 180] loss: 0.772 [20, 240] loss: 0.747 [20, 300] loss: 0.762 [20, 360] loss: 0.758 Epoch: 20 -> Loss: 0.672998905182 Epoch: 20 -> Test Accuracy: 68.56 [21, 60] loss: 0.716 [21, 120] loss: 0.697 [21, 180] loss: 0.697 [21, 240] loss: 0.668 [21, 300] loss: 0.656 [21, 360] loss: 0.666 Epoch: 21 -> Loss: 0.708767354488 Epoch: 21 -> Test Accuracy: 70.52 [22, 60] loss: 0.682 [22, 120] loss: 0.660 [22, 180] loss: 0.658 [22, 240] loss: 0.655 [22, 300] loss: 0.644 [22, 360] loss: 0.642 Epoch: 22 -> Loss: 0.594831764698 Epoch: 22 -> Test Accuracy: 71.68 [23, 60] loss: 0.632 [23, 120] loss: 0.630 [23, 180] loss: 0.635 [23, 240] loss: 0.637 [23, 300] loss: 0.631 [23, 360] loss: 0.652 Epoch: 23 -> Loss: 0.603423058987 Epoch: 23 -> Test Accuracy: 71.55 [24, 60] loss: 0.632 [24, 120] loss: 0.620 [24, 180] loss: 0.627 [24, 240] loss: 0.623 [24, 300] loss: 0.630 [24, 360] loss: 0.639 Epoch: 24 -> Loss: 0.603372693062 Epoch: 24 -> Test Accuracy: 71.9 [25, 60] loss: 0.623 [25, 120] loss: 0.628 [25, 180] loss: 0.639 [25, 240] loss: 0.601 [25, 300] loss: 0.632 [25, 360] loss: 0.626 Epoch: 25 -> Loss: 0.616600453854 Epoch: 25 -> Test Accuracy: 71.63 [26, 60] loss: 0.609 [26, 120] loss: 0.640 [26, 180] loss: 0.600 [26, 240] loss: 0.610 [26, 300] loss: 0.592 [26, 360] loss: 0.637 Epoch: 26 -> Loss: 0.608256340027 Epoch: 26 -> Test Accuracy: 72.46 [27, 60] loss: 0.608 [27, 120] loss: 0.624 [27, 180] loss: 0.601 [27, 240] loss: 0.627 [27, 300] loss: 0.604 [27, 360] loss: 0.643 Epoch: 27 -> Loss: 0.691409289837 Epoch: 27 -> Test Accuracy: 71.53 [28, 60] loss: 0.615 [28, 120] loss: 0.607 [28, 180] loss: 0.612 [28, 240] loss: 0.630 [28, 300] loss: 0.616 [28, 360] loss: 0.615 Epoch: 28 -> Loss: 0.613882958889 Epoch: 28 -> Test Accuracy: 72.31 [29, 60] loss: 0.611 [29, 120] loss: 0.606 [29, 180] loss: 0.625 [29, 240] loss: 0.623 [29, 300] loss: 0.606 [29, 360] loss: 0.616 Epoch: 29 -> Loss: 0.467459022999 Epoch: 29 -> Test Accuracy: 72.25 [30, 60] loss: 0.610 [30, 120] loss: 0.625 [30, 180] loss: 0.606 [30, 240] loss: 0.618 [30, 300] loss: 0.610 [30, 360] loss: 0.614 Epoch: 30 -> Loss: 0.833746314049 Epoch: 30 -> Test Accuracy: 72.45 [31, 60] loss: 0.607 [31, 120] loss: 0.608 [31, 180] loss: 0.630 [31, 240] loss: 0.606 [31, 300] loss: 0.614 [31, 360] loss: 0.596 Epoch: 31 -> Loss: 0.659976422787 Epoch: 31 -> Test Accuracy: 71.82 [32, 60] loss: 0.614 [32, 120] loss: 0.597 [32, 180] loss: 0.616 [32, 240] loss: 0.601 [32, 300] loss: 0.607 [32, 360] loss: 0.623 Epoch: 32 -> Loss: 0.54079246521 Epoch: 32 -> Test Accuracy: 71.52 [33, 60] loss: 0.603 [33, 120] loss: 0.626 [33, 180] loss: 0.604 [33, 240] loss: 0.611 [33, 300] loss: 0.583 [33, 360] loss: 0.614 Epoch: 33 -> Loss: 0.495990037918 Epoch: 33 -> Test Accuracy: 71.8 [34, 60] loss: 0.600 [34, 120] loss: 0.601 [34, 180] loss: 0.596 [34, 240] loss: 0.602 [34, 300] loss: 0.620 [34, 360] loss: 0.600 Epoch: 34 -> Loss: 0.66494768858 Epoch: 34 -> Test Accuracy: 71.88 [35, 60] loss: 0.595 [35, 120] loss: 0.617 [35, 180] loss: 0.625 [35, 240] loss: 0.607 [35, 300] loss: 0.590 [35, 360] loss: 0.622 Epoch: 35 -> Loss: 0.629398822784 Epoch: 35 -> Test Accuracy: 72.17 [36, 60] loss: 0.609 [36, 120] loss: 0.584 [36, 180] loss: 0.606 [36, 240] loss: 0.605 [36, 300] loss: 0.616 [36, 360] loss: 0.607 Epoch: 36 -> Loss: 0.558319091797 Epoch: 36 -> Test Accuracy: 72.36 [37, 60] loss: 0.599 [37, 120] loss: 0.606 [37, 180] loss: 0.601 [37, 240] loss: 0.600 [37, 300] loss: 0.603 [37, 360] loss: 0.612 Epoch: 37 -> Loss: 0.560134530067 Epoch: 37 -> Test Accuracy: 72.43 [38, 60] loss: 0.590 [38, 120] loss: 0.613 [38, 180] loss: 0.617 [38, 240] loss: 0.618 [38, 300] loss: 0.603 [38, 360] loss: 0.603 Epoch: 38 -> Loss: 0.708665013313 Epoch: 38 -> Test Accuracy: 72.11 [39, 60] loss: 0.603 [39, 120] loss: 0.614 [39, 180] loss: 0.598 [39, 240] loss: 0.612 [39, 300] loss: 0.615 [39, 360] loss: 0.606 Epoch: 39 -> Loss: 0.652741909027 Epoch: 39 -> Test Accuracy: 72.6 [40, 60] loss: 0.580 [40, 120] loss: 0.595 [40, 180] loss: 0.591 [40, 240] loss: 0.612 [40, 300] loss: 0.604 [40, 360] loss: 0.622 Epoch: 40 -> Loss: 0.494285404682 Epoch: 40 -> Test Accuracy: 71.79 [41, 60] loss: 0.571 [41, 120] loss: 0.574 [41, 180] loss: 0.549 [41, 240] loss: 0.574 [41, 300] loss: 0.553 [41, 360] loss: 0.548 Epoch: 41 -> Loss: 0.537647306919 Epoch: 41 -> Test Accuracy: 73.24 [42, 60] loss: 0.551 [42, 120] loss: 0.557 [42, 180] loss: 0.538 [42, 240] loss: 0.547 [42, 300] loss: 0.523 [42, 360] loss: 0.536 Epoch: 42 -> Loss: 0.584192574024 Epoch: 42 -> Test Accuracy: 73.78 [43, 60] loss: 0.537 [43, 120] loss: 0.525 [43, 180] loss: 0.539 [43, 240] loss: 0.546 [43, 300] loss: 0.525 [43, 360] loss: 0.535 Epoch: 43 -> Loss: 0.75417226553 Epoch: 43 -> Test Accuracy: 73.71 [44, 60] loss: 0.526 [44, 120] loss: 0.518 [44, 180] loss: 0.523 [44, 240] loss: 0.527 [44, 300] loss: 0.512 [44, 360] loss: 0.518 Epoch: 44 -> Loss: 0.419095039368 Epoch: 44 -> Test Accuracy: 74.0 [45, 60] loss: 0.522 [45, 120] loss: 0.509 [45, 180] loss: 0.497 [45, 240] loss: 0.524 [45, 300] loss: 0.527 [45, 360] loss: 0.521 Epoch: 45 -> Loss: 0.390105068684 Epoch: 45 -> Test Accuracy: 74.02 [46, 60] loss: 0.506 [46, 120] loss: 0.499 [46, 180] loss: 0.513 [46, 240] loss: 0.515 [46, 300] loss: 0.514 [46, 360] loss: 0.508 Epoch: 46 -> Loss: 0.426923751831 Epoch: 46 -> Test Accuracy: 74.18 [47, 60] loss: 0.535 [47, 120] loss: 0.511 [47, 180] loss: 0.503 [47, 240] loss: 0.537 [47, 300] loss: 0.500 [47, 360] loss: 0.499 Epoch: 47 -> Loss: 0.380633890629 Epoch: 47 -> Test Accuracy: 74.3 [48, 60] loss: 0.519 [48, 120] loss: 0.508 [48, 180] loss: 0.500 [48, 240] loss: 0.499 [48, 300] loss: 0.483 [48, 360] loss: 0.486 Epoch: 48 -> Loss: 0.623250901699 Epoch: 48 -> Test Accuracy: 74.36 [49, 60] loss: 0.497 [49, 120] loss: 0.492 [49, 180] loss: 0.505 [49, 240] loss: 0.502 [49, 300] loss: 0.508 [49, 360] loss: 0.506 Epoch: 49 -> Loss: 0.539638578892 Epoch: 49 -> Test Accuracy: 74.28 [50, 60] loss: 0.494 [50, 120] loss: 0.513 [50, 180] loss: 0.495 [50, 240] loss: 0.500 [50, 300] loss: 0.507 [50, 360] loss: 0.500 Epoch: 50 -> Loss: 0.487371295691 Epoch: 50 -> Test Accuracy: 74.31 [51, 60] loss: 0.493 [51, 120] loss: 0.494 [51, 180] loss: 0.491 [51, 240] loss: 0.507 [51, 300] loss: 0.498 [51, 360] loss: 0.503 Epoch: 51 -> Loss: 0.542852222919 Epoch: 51 -> Test Accuracy: 74.13 [52, 60] loss: 0.485 [52, 120] loss: 0.497 [52, 180] loss: 0.504 [52, 240] loss: 0.504 [52, 300] loss: 0.498 [52, 360] loss: 0.481 Epoch: 52 -> Loss: 0.670035123825 Epoch: 52 -> Test Accuracy: 74.38 [53, 60] loss: 0.499 [53, 120] loss: 0.497 [53, 180] loss: 0.485 [53, 240] loss: 0.502 [53, 300] loss: 0.504 [53, 360] loss: 0.507 Epoch: 53 -> Loss: 0.512040674686 Epoch: 53 -> Test Accuracy: 74.27 [54, 60] loss: 0.492 [54, 120] loss: 0.498 [54, 180] loss: 0.477 [54, 240] loss: 0.486 [54, 300] loss: 0.483 [54, 360] loss: 0.491 Epoch: 54 -> Loss: 0.403091520071 Epoch: 54 -> Test Accuracy: 74.45 [55, 60] loss: 0.494 [55, 120] loss: 0.498 [55, 180] loss: 0.493 [55, 240] loss: 0.507 [55, 300] loss: 0.497 [55, 360] loss: 0.496 Epoch: 55 -> Loss: 0.333195716143 Epoch: 55 -> Test Accuracy: 74.37 [56, 60] loss: 0.488 [56, 120] loss: 0.479 [56, 180] loss: 0.512 [56, 240] loss: 0.490 [56, 300] loss: 0.492 [56, 360] loss: 0.492 Epoch: 56 -> Loss: 0.65333122015 Epoch: 56 -> Test Accuracy: 74.53 [57, 60] loss: 0.484 [57, 120] loss: 0.497 [57, 180] loss: 0.489 [57, 240] loss: 0.481 [57, 300] loss: 0.483 [57, 360] loss: 0.495 Epoch: 57 -> Loss: 0.746136724949 Epoch: 57 -> Test Accuracy: 74.45 [58, 60] loss: 0.492 [58, 120] loss: 0.505 [58, 180] loss: 0.515 [58, 240] loss: 0.496 [58, 300] loss: 0.484 [58, 360] loss: 0.482 Epoch: 58 -> Loss: 0.403785765171 Epoch: 58 -> Test Accuracy: 74.54 [59, 60] loss: 0.481 [59, 120] loss: 0.491 [59, 180] loss: 0.490 [59, 240] loss: 0.484 [59, 300] loss: 0.484 [59, 360] loss: 0.488 Epoch: 59 -> Loss: 0.374584436417 Epoch: 59 -> Test Accuracy: 74.79 [60, 60] loss: 0.481 [60, 120] loss: 0.487 [60, 180] loss: 0.483 [60, 240] loss: 0.501 [60, 300] loss: 0.494 [60, 360] loss: 0.495 Epoch: 60 -> Loss: 0.36842250824 Epoch: 60 -> Test Accuracy: 74.52 [61, 60] loss: 0.490 [61, 120] loss: 0.479 [61, 180] loss: 0.492 [61, 240] loss: 0.480 [61, 300] loss: 0.478 [61, 360] loss: 0.481 Epoch: 61 -> Loss: 0.533181369305 Epoch: 61 -> Test Accuracy: 74.47 [62, 60] loss: 0.502 [62, 120] loss: 0.470 [62, 180] loss: 0.501 [62, 240] loss: 0.480 [62, 300] loss: 0.473 [62, 360] loss: 0.469 Epoch: 62 -> Loss: 0.488505065441 Epoch: 62 -> Test Accuracy: 74.44 [63, 60] loss: 0.489 [63, 120] loss: 0.482 [63, 180] loss: 0.486 [63, 240] loss: 0.479 [63, 300] loss: 0.492 [63, 360] loss: 0.496 Epoch: 63 -> Loss: 0.505281805992 Epoch: 63 -> Test Accuracy: 74.81 [64, 60] loss: 0.480 [64, 120] loss: 0.490 [64, 180] loss: 0.479 [64, 240] loss: 0.482 [64, 300] loss: 0.487 [64, 360] loss: 0.496 Epoch: 64 -> Loss: 0.606616079807 Epoch: 64 -> Test Accuracy: 74.7 [65, 60] loss: 0.500 [65, 120] loss: 0.480 [65, 180] loss: 0.470 [65, 240] loss: 0.479 [65, 300] loss: 0.487 [65, 360] loss: 0.482 Epoch: 65 -> Loss: 0.629588782787 Epoch: 65 -> Test Accuracy: 74.61 [66, 60] loss: 0.485 [66, 120] loss: 0.495 [66, 180] loss: 0.478 [66, 240] loss: 0.487 [66, 300] loss: 0.469 [66, 360] loss: 0.474 Epoch: 66 -> Loss: 0.429427206516 Epoch: 66 -> Test Accuracy: 74.81 [67, 60] loss: 0.468 [67, 120] loss: 0.478 [67, 180] loss: 0.495 [67, 240] loss: 0.491 [67, 300] loss: 0.486 [67, 360] loss: 0.476 Epoch: 67 -> Loss: 0.418888032436 Epoch: 67 -> Test Accuracy: 74.74 [68, 60] loss: 0.463 [68, 120] loss: 0.465 [68, 180] loss: 0.494 [68, 240] loss: 0.484 [68, 300] loss: 0.475 [68, 360] loss: 0.511 Epoch: 68 -> Loss: 0.612156569958 Epoch: 68 -> Test Accuracy: 74.77 [69, 60] loss: 0.478 [69, 120] loss: 0.483 [69, 180] loss: 0.483 [69, 240] loss: 0.482 [69, 300] loss: 0.476 [69, 360] loss: 0.468 Epoch: 69 -> Loss: 0.6005885005 Epoch: 69 -> Test Accuracy: 74.73 [70, 60] loss: 0.464 [70, 120] loss: 0.493 [70, 180] loss: 0.470 [70, 240] loss: 0.478 [70, 300] loss: 0.482 [70, 360] loss: 0.479 Epoch: 70 -> Loss: 0.673352837563 Epoch: 70 -> Test Accuracy: 74.79 [71, 60] loss: 0.485 [71, 120] loss: 0.491 [71, 180] loss: 0.478 [71, 240] loss: 0.480 [71, 300] loss: 0.457 [71, 360] loss: 0.476 Epoch: 71 -> Loss: 0.438875764608 Epoch: 71 -> Test Accuracy: 74.59 [72, 60] loss: 0.487 [72, 120] loss: 0.471 [72, 180] loss: 0.476 [72, 240] loss: 0.456 [72, 300] loss: 0.474 [72, 360] loss: 0.488 Epoch: 72 -> Loss: 0.525793015957 Epoch: 72 -> Test Accuracy: 74.68 [73, 60] loss: 0.465 [73, 120] loss: 0.482 [73, 180] loss: 0.466 [73, 240] loss: 0.487 [73, 300] loss: 0.474 [73, 360] loss: 0.505 Epoch: 73 -> Loss: 0.405533373356 Epoch: 73 -> Test Accuracy: 74.89 [74, 60] loss: 0.474 [74, 120] loss: 0.475 [74, 180] loss: 0.464 [74, 240] loss: 0.488 [74, 300] loss: 0.484 [74, 360] loss: 0.474 Epoch: 74 -> Loss: 0.626014411449 Epoch: 74 -> Test Accuracy: 74.86 [75, 60] loss: 0.481 [75, 120] loss: 0.479 [75, 180] loss: 0.477 [75, 240] loss: 0.473 [75, 300] loss: 0.472 [75, 360] loss: 0.467 Epoch: 75 -> Loss: 0.482714742422 Epoch: 75 -> Test Accuracy: 74.84 [76, 60] loss: 0.481 [76, 120] loss: 0.450 [76, 180] loss: 0.480 [76, 240] loss: 0.480 [76, 300] loss: 0.465 [76, 360] loss: 0.469 Epoch: 76 -> Loss: 0.434392929077 Epoch: 76 -> Test Accuracy: 74.81 [77, 60] loss: 0.458 [77, 120] loss: 0.473 [77, 180] loss: 0.478 [77, 240] loss: 0.457 [77, 300] loss: 0.487 [77, 360] loss: 0.489 Epoch: 77 -> Loss: 0.608023047447 Epoch: 77 -> Test Accuracy: 74.88 [78, 60] loss: 0.473 [78, 120] loss: 0.476 [78, 180] loss: 0.468 [78, 240] loss: 0.469 [78, 300] loss: 0.460 [78, 360] loss: 0.487 Epoch: 78 -> Loss: 0.561837613583 Epoch: 78 -> Test Accuracy: 74.88 [79, 60] loss: 0.482 [79, 120] loss: 0.479 [79, 180] loss: 0.464 [79, 240] loss: 0.475 [79, 300] loss: 0.472 [79, 360] loss: 0.481 Epoch: 79 -> Loss: 0.610006928444 Epoch: 79 -> Test Accuracy: 74.86 [80, 60] loss: 0.472 [80, 120] loss: 0.470 [80, 180] loss: 0.476 [80, 240] loss: 0.470 [80, 300] loss: 0.481 [80, 360] loss: 0.469 Epoch: 80 -> Loss: 0.624663233757 Epoch: 80 -> Test Accuracy: 74.8 [81, 60] loss: 0.481 [81, 120] loss: 0.466 [81, 180] loss: 0.472 [81, 240] loss: 0.473 [81, 300] loss: 0.463 [81, 360] loss: 0.460 Epoch: 81 -> Loss: 0.552075564861 Epoch: 81 -> Test Accuracy: 74.83 [82, 60] loss: 0.465 [82, 120] loss: 0.464 [82, 180] loss: 0.473 [82, 240] loss: 0.453 [82, 300] loss: 0.473 [82, 360] loss: 0.473 Epoch: 82 -> Loss: 0.435061633587 Epoch: 82 -> Test Accuracy: 74.76 [83, 60] loss: 0.464 [83, 120] loss: 0.475 [83, 180] loss: 0.471 [83, 240] loss: 0.466 [83, 300] loss: 0.478 [83, 360] loss: 0.457 Epoch: 83 -> Loss: 0.518538594246 Epoch: 83 -> Test Accuracy: 74.86 [84, 60] loss: 0.456 [84, 120] loss: 0.467 [84, 180] loss: 0.472 [84, 240] loss: 0.484 [84, 300] loss: 0.481 [84, 360] loss: 0.452 Epoch: 84 -> Loss: 0.371880471706 Epoch: 84 -> Test Accuracy: 74.77 [85, 60] loss: 0.455 [85, 120] loss: 0.456 [85, 180] loss: 0.477 [85, 240] loss: 0.476 [85, 300] loss: 0.468 [85, 360] loss: 0.469 Epoch: 85 -> Loss: 0.559426724911 Epoch: 85 -> Test Accuracy: 74.75 [86, 60] loss: 0.477 [86, 120] loss: 0.465 [86, 180] loss: 0.481 [86, 240] loss: 0.468 [86, 300] loss: 0.468 [86, 360] loss: 0.463 Epoch: 86 -> Loss: 0.523461937904 Epoch: 86 -> Test Accuracy: 74.9 [87, 60] loss: 0.462 [87, 120] loss: 0.458 [87, 180] loss: 0.466 [87, 240] loss: 0.464 [87, 300] loss: 0.476 [87, 360] loss: 0.475 Epoch: 87 -> Loss: 0.357460737228 Epoch: 87 -> Test Accuracy: 75.07 [88, 60] loss: 0.452 [88, 120] loss: 0.470 [88, 180] loss: 0.475 [88, 240] loss: 0.445 [88, 300] loss: 0.478 [88, 360] loss: 0.466 Epoch: 88 -> Loss: 0.404544979334 Epoch: 88 -> Test Accuracy: 75.1 [89, 60] loss: 0.451 [89, 120] loss: 0.458 [89, 180] loss: 0.456 [89, 240] loss: 0.477 [89, 300] loss: 0.466 [89, 360] loss: 0.458 Epoch: 89 -> Loss: 0.439839839935 Epoch: 89 -> Test Accuracy: 74.95 [90, 60] loss: 0.464 [90, 120] loss: 0.452 [90, 180] loss: 0.471 [90, 240] loss: 0.464 [90, 300] loss: 0.457 [90, 360] loss: 0.480 Epoch: 90 -> Loss: 0.471448481083 Epoch: 90 -> Test Accuracy: 74.75 [91, 60] loss: 0.452 [91, 120] loss: 0.468 [91, 180] loss: 0.457 [91, 240] loss: 0.483 [91, 300] loss: 0.457 [91, 360] loss: 0.464 Epoch: 91 -> Loss: 0.596991837025 Epoch: 91 -> Test Accuracy: 74.89 [92, 60] loss: 0.451 [92, 120] loss: 0.462 [92, 180] loss: 0.453 [92, 240] loss: 0.458 [92, 300] loss: 0.472 [92, 360] loss: 0.486 Epoch: 92 -> Loss: 0.657341241837 Epoch: 92 -> Test Accuracy: 74.9 [93, 60] loss: 0.455 [93, 120] loss: 0.452 [93, 180] loss: 0.461 [93, 240] loss: 0.465 [93, 300] loss: 0.442 [93, 360] loss: 0.455 Epoch: 93 -> Loss: 0.535405635834 Epoch: 93 -> Test Accuracy: 74.8 [94, 60] loss: 0.460 [94, 120] loss: 0.466 [94, 180] loss: 0.468 [94, 240] loss: 0.475 [94, 300] loss: 0.450 [94, 360] loss: 0.460 Epoch: 94 -> Loss: 0.549831449986 Epoch: 94 -> Test Accuracy: 74.98 [95, 60] loss: 0.458 [95, 120] loss: 0.462 [95, 180] loss: 0.468 [95, 240] loss: 0.468 [95, 300] loss: 0.461 [95, 360] loss: 0.463 Epoch: 95 -> Loss: 0.603649318218 Epoch: 95 -> Test Accuracy: 75.06 [96, 60] loss: 0.453 [96, 120] loss: 0.454 [96, 180] loss: 0.466 [96, 240] loss: 0.456 [96, 300] loss: 0.453 [96, 360] loss: 0.458 Epoch: 96 -> Loss: 0.593458533287 Epoch: 96 -> Test Accuracy: 74.84 [97, 60] loss: 0.451 [97, 120] loss: 0.475 [97, 180] loss: 0.452 [97, 240] loss: 0.461 [97, 300] loss: 0.458 [97, 360] loss: 0.451 Epoch: 97 -> Loss: 0.493361532688 Epoch: 97 -> Test Accuracy: 75.17 [98, 60] loss: 0.450 [98, 120] loss: 0.456 [98, 180] loss: 0.473 [98, 240] loss: 0.464 [98, 300] loss: 0.466 [98, 360] loss: 0.458 Epoch: 98 -> Loss: 0.489044964314 Epoch: 98 -> Test Accuracy: 74.98 [99, 60] loss: 0.455 [99, 120] loss: 0.455 [99, 180] loss: 0.449 [99, 240] loss: 0.450 [99, 300] loss: 0.458 [99, 360] loss: 0.468 Epoch: 99 -> Loss: 0.518808782101 Epoch: 99 -> Test Accuracy: 74.67 [100, 60] loss: 0.442 [100, 120] loss: 0.463 [100, 180] loss: 0.458 [100, 240] loss: 0.453 [100, 300] loss: 0.453 [100, 360] loss: 0.461 Epoch: 100 -> Loss: 0.443654119968 Epoch: 100 -> Test Accuracy: 75.13 Finished Training [1, 60] loss: 2.802 [1, 120] loss: 2.132 [1, 180] loss: 2.093 [1, 240] loss: 2.042 [1, 300] loss: 2.024 [1, 360] loss: 1.988 Epoch: 1 -> Loss: 1.84517610073 Epoch: 1 -> Test Accuracy: 26.21 [2, 60] loss: 1.976 [2, 120] loss: 1.960 [2, 180] loss: 1.948 [2, 240] loss: 1.929 [2, 300] loss: 1.928 [2, 360] loss: 1.913 Epoch: 2 -> Loss: 1.92097055912 Epoch: 2 -> Test Accuracy: 27.91 [3, 60] loss: 1.906 [3, 120] loss: 1.906 [3, 180] loss: 1.900 [3, 240] loss: 1.878 [3, 300] loss: 1.884 [3, 360] loss: 1.867 Epoch: 3 -> Loss: 1.91858863831 Epoch: 3 -> Test Accuracy: 28.33 [4, 60] loss: 1.874 [4, 120] loss: 1.874 [4, 180] loss: 1.866 [4, 240] loss: 1.871 [4, 300] loss: 1.846 [4, 360] loss: 1.847 Epoch: 4 -> Loss: 1.81260871887 Epoch: 4 -> Test Accuracy: 30.25 [5, 60] loss: 1.851 [5, 120] loss: 1.842 [5, 180] loss: 1.848 [5, 240] loss: 1.845 [5, 300] loss: 1.848 [5, 360] loss: 1.840 Epoch: 5 -> Loss: 2.05401945114 Epoch: 5 -> Test Accuracy: 30.04 [6, 60] loss: 1.833 [6, 120] loss: 1.826 [6, 180] loss: 1.848 [6, 240] loss: 1.822 [6, 300] loss: 1.825 [6, 360] loss: 1.809 Epoch: 6 -> Loss: 1.8115953207 Epoch: 6 -> Test Accuracy: 30.62 [7, 60] loss: 1.824 [7, 120] loss: 1.817 [7, 180] loss: 1.845 [7, 240] loss: 1.836 [7, 300] loss: 1.814 [7, 360] loss: 1.828 Epoch: 7 -> Loss: 2.0414352417 Epoch: 7 -> Test Accuracy: 31.99 [8, 60] loss: 1.813 [8, 120] loss: 1.830 [8, 180] loss: 1.834 [8, 240] loss: 1.820 [8, 300] loss: 1.821 [8, 360] loss: 1.804 Epoch: 8 -> Loss: 1.76224195957 Epoch: 8 -> Test Accuracy: 31.04 [9, 60] loss: 1.798 [9, 120] loss: 1.803 [9, 180] loss: 1.819 [9, 240] loss: 1.829 [9, 300] loss: 1.822 [9, 360] loss: 1.813 Epoch: 9 -> Loss: 1.81824815273 Epoch: 9 -> Test Accuracy: 31.84 [10, 60] loss: 1.808 [10, 120] loss: 1.823 [10, 180] loss: 1.811 [10, 240] loss: 1.820 [10, 300] loss: 1.803 [10, 360] loss: 1.811 Epoch: 10 -> Loss: 1.829033494 Epoch: 10 -> Test Accuracy: 31.31 [11, 60] loss: 1.784 [11, 120] loss: 1.824 [11, 180] loss: 1.803 [11, 240] loss: 1.802 [11, 300] loss: 1.810 [11, 360] loss: 1.793 Epoch: 11 -> Loss: 1.78773093224 Epoch: 11 -> Test Accuracy: 31.66 [12, 60] loss: 1.809 [12, 120] loss: 1.795 [12, 180] loss: 1.796 [12, 240] loss: 1.796 [12, 300] loss: 1.793 [12, 360] loss: 1.795 Epoch: 12 -> Loss: 1.76458454132 Epoch: 12 -> Test Accuracy: 30.73 [13, 60] loss: 1.822 [13, 120] loss: 1.784 [13, 180] loss: 1.799 [13, 240] loss: 1.794 [13, 300] loss: 1.798 [13, 360] loss: 1.783 Epoch: 13 -> Loss: 1.91414809227 Epoch: 13 -> Test Accuracy: 31.58 [14, 60] loss: 1.793 [14, 120] loss: 1.793 [14, 180] loss: 1.796 [14, 240] loss: 1.792 [14, 300] loss: 1.795 [14, 360] loss: 1.780 Epoch: 14 -> Loss: 1.82072198391 Epoch: 14 -> Test Accuracy: 31.69 [15, 60] loss: 1.797 [15, 120] loss: 1.790 [15, 180] loss: 1.784 [15, 240] loss: 1.796 [15, 300] loss: 1.791 [15, 360] loss: 1.800 Epoch: 15 -> Loss: 1.71568238735 Epoch: 15 -> Test Accuracy: 31.9 [16, 60] loss: 1.802 [16, 120] loss: 1.797 [16, 180] loss: 1.772 [16, 240] loss: 1.789 [16, 300] loss: 1.797 [16, 360] loss: 1.782 Epoch: 16 -> Loss: 1.79029119015 Epoch: 16 -> Test Accuracy: 31.96 [17, 60] loss: 1.788 [17, 120] loss: 1.782 [17, 180] loss: 1.783 [17, 240] loss: 1.805 [17, 300] loss: 1.791 [17, 360] loss: 1.780 Epoch: 17 -> Loss: 1.95210969448 Epoch: 17 -> Test Accuracy: 32.56 [18, 60] loss: 1.796 [18, 120] loss: 1.784 [18, 180] loss: 1.779 [18, 240] loss: 1.807 [18, 300] loss: 1.804 [18, 360] loss: 1.787 Epoch: 18 -> Loss: 1.68113386631 Epoch: 18 -> Test Accuracy: 32.32 [19, 60] loss: 1.782 [19, 120] loss: 1.792 [19, 180] loss: 1.789 [19, 240] loss: 1.785 [19, 300] loss: 1.777 [19, 360] loss: 1.794 Epoch: 19 -> Loss: 1.90905153751 Epoch: 19 -> Test Accuracy: 33.03 [20, 60] loss: 1.771 [20, 120] loss: 1.779 [20, 180] loss: 1.766 [20, 240] loss: 1.790 [20, 300] loss: 1.783 [20, 360] loss: 1.792 Epoch: 20 -> Loss: 1.83599281311 Epoch: 20 -> Test Accuracy: 32.22 [21, 60] loss: 1.736 [21, 120] loss: 1.718 [21, 180] loss: 1.704 [21, 240] loss: 1.719 [21, 300] loss: 1.697 [21, 360] loss: 1.698 Epoch: 21 -> Loss: 1.66823065281 Epoch: 21 -> Test Accuracy: 34.98 [22, 60] loss: 1.688 [22, 120] loss: 1.687 [22, 180] loss: 1.680 [22, 240] loss: 1.692 [22, 300] loss: 1.670 [22, 360] loss: 1.679 Epoch: 22 -> Loss: 1.74744164944 Epoch: 22 -> Test Accuracy: 35.23 [23, 60] loss: 1.665 [23, 120] loss: 1.687 [23, 180] loss: 1.670 [23, 240] loss: 1.674 [23, 300] loss: 1.658 [23, 360] loss: 1.690 Epoch: 23 -> Loss: 1.64546132088 Epoch: 23 -> Test Accuracy: 34.83 [24, 60] loss: 1.677 [24, 120] loss: 1.676 [24, 180] loss: 1.672 [24, 240] loss: 1.674 [24, 300] loss: 1.667 [24, 360] loss: 1.666 Epoch: 24 -> Loss: 1.57270264626 Epoch: 24 -> Test Accuracy: 35.0 [25, 60] loss: 1.675 [25, 120] loss: 1.666 [25, 180] loss: 1.668 [25, 240] loss: 1.655 [25, 300] loss: 1.664 [25, 360] loss: 1.669 Epoch: 25 -> Loss: 1.66615843773 Epoch: 25 -> Test Accuracy: 35.75 [26, 60] loss: 1.652 [26, 120] loss: 1.671 [26, 180] loss: 1.662 [26, 240] loss: 1.651 [26, 300] loss: 1.678 [26, 360] loss: 1.654 Epoch: 26 -> Loss: 1.68612635136 Epoch: 26 -> Test Accuracy: 35.25 [27, 60] loss: 1.668 [27, 120] loss: 1.662 [27, 180] loss: 1.663 [27, 240] loss: 1.645 [27, 300] loss: 1.653 [27, 360] loss: 1.679 Epoch: 27 -> Loss: 1.61806702614 Epoch: 27 -> Test Accuracy: 35.38 [28, 60] loss: 1.649 [28, 120] loss: 1.631 [28, 180] loss: 1.669 [28, 240] loss: 1.662 [28, 300] loss: 1.671 [28, 360] loss: 1.670 Epoch: 28 -> Loss: 1.65237653255 Epoch: 28 -> Test Accuracy: 35.88 [29, 60] loss: 1.657 [29, 120] loss: 1.661 [29, 180] loss: 1.666 [29, 240] loss: 1.649 [29, 300] loss: 1.648 [29, 360] loss: 1.645 Epoch: 29 -> Loss: 1.72619855404 Epoch: 29 -> Test Accuracy: 35.23 [30, 60] loss: 1.676 [30, 120] loss: 1.655 [30, 180] loss: 1.647 [30, 240] loss: 1.642 [30, 300] loss: 1.650 [30, 360] loss: 1.660 Epoch: 30 -> Loss: 1.58070516586 Epoch: 30 -> Test Accuracy: 35.45 [31, 60] loss: 1.658 [31, 120] loss: 1.620 [31, 180] loss: 1.655 [31, 240] loss: 1.662 [31, 300] loss: 1.651 [31, 360] loss: 1.659 Epoch: 31 -> Loss: 1.78566479683 Epoch: 31 -> Test Accuracy: 36.28 [32, 60] loss: 1.642 [32, 120] loss: 1.647 [32, 180] loss: 1.651 [32, 240] loss: 1.661 [32, 300] loss: 1.657 [32, 360] loss: 1.641 Epoch: 32 -> Loss: 1.61401367188 Epoch: 32 -> Test Accuracy: 35.42 [33, 60] loss: 1.628 [33, 120] loss: 1.667 [33, 180] loss: 1.657 [33, 240] loss: 1.666 [33, 300] loss: 1.646 [33, 360] loss: 1.651 Epoch: 33 -> Loss: 1.46038353443 Epoch: 33 -> Test Accuracy: 36.32 [34, 60] loss: 1.650 [34, 120] loss: 1.639 [34, 180] loss: 1.663 [34, 240] loss: 1.631 [34, 300] loss: 1.661 [34, 360] loss: 1.664 Epoch: 34 -> Loss: 1.63921451569 Epoch: 34 -> Test Accuracy: 36.24 [35, 60] loss: 1.637 [35, 120] loss: 1.658 [35, 180] loss: 1.670 [35, 240] loss: 1.641 [35, 300] loss: 1.654 [35, 360] loss: 1.638 Epoch: 35 -> Loss: 1.52031636238 Epoch: 35 -> Test Accuracy: 36.75 [36, 60] loss: 1.653 [36, 120] loss: 1.652 [36, 180] loss: 1.650 [36, 240] loss: 1.651 [36, 300] loss: 1.656 [36, 360] loss: 1.642 Epoch: 36 -> Loss: 1.73019337654 Epoch: 36 -> Test Accuracy: 35.95 [37, 60] loss: 1.640 [37, 120] loss: 1.655 [37, 180] loss: 1.636 [37, 240] loss: 1.650 [37, 300] loss: 1.634 [37, 360] loss: 1.662 Epoch: 37 -> Loss: 1.7520275116 Epoch: 37 -> Test Accuracy: 35.59 [38, 60] loss: 1.642 [38, 120] loss: 1.650 [38, 180] loss: 1.643 [38, 240] loss: 1.656 [38, 300] loss: 1.664 [38, 360] loss: 1.661 Epoch: 38 -> Loss: 1.50345671177 Epoch: 38 -> Test Accuracy: 35.78 [39, 60] loss: 1.621 [39, 120] loss: 1.617 [39, 180] loss: 1.660 [39, 240] loss: 1.665 [39, 300] loss: 1.663 [39, 360] loss: 1.644 Epoch: 39 -> Loss: 1.67437672615 Epoch: 39 -> Test Accuracy: 36.05 [40, 60] loss: 1.637 [40, 120] loss: 1.635 [40, 180] loss: 1.636 [40, 240] loss: 1.656 [40, 300] loss: 1.654 [40, 360] loss: 1.649 Epoch: 40 -> Loss: 1.69791388512 Epoch: 40 -> Test Accuracy: 35.94 [41, 60] loss: 1.631 [41, 120] loss: 1.613 [41, 180] loss: 1.611 [41, 240] loss: 1.585 [41, 300] loss: 1.596 [41, 360] loss: 1.597 Epoch: 41 -> Loss: 1.52905249596 Epoch: 41 -> Test Accuracy: 37.44 [42, 60] loss: 1.589 [42, 120] loss: 1.605 [42, 180] loss: 1.592 [42, 240] loss: 1.599 [42, 300] loss: 1.600 [42, 360] loss: 1.590 Epoch: 42 -> Loss: 1.51411664486 Epoch: 42 -> Test Accuracy: 37.54 [43, 60] loss: 1.576 [43, 120] loss: 1.552 [43, 180] loss: 1.592 [43, 240] loss: 1.590 [43, 300] loss: 1.566 [43, 360] loss: 1.588 Epoch: 43 -> Loss: 1.48796474934 Epoch: 43 -> Test Accuracy: 37.68 [44, 60] loss: 1.589 [44, 120] loss: 1.567 [44, 180] loss: 1.567 [44, 240] loss: 1.574 [44, 300] loss: 1.585 [44, 360] loss: 1.566 Epoch: 44 -> Loss: 1.55248272419 Epoch: 44 -> Test Accuracy: 37.51 [45, 60] loss: 1.573 [45, 120] loss: 1.583 [45, 180] loss: 1.589 [45, 240] loss: 1.558 [45, 300] loss: 1.576 [45, 360] loss: 1.584 Epoch: 45 -> Loss: 1.66815686226 Epoch: 45 -> Test Accuracy: 37.7 [46, 60] loss: 1.546 [46, 120] loss: 1.581 [46, 180] loss: 1.573 [46, 240] loss: 1.554 [46, 300] loss: 1.549 [46, 360] loss: 1.562 Epoch: 46 -> Loss: 1.74636232853 Epoch: 46 -> Test Accuracy: 38.27 [47, 60] loss: 1.559 [47, 120] loss: 1.578 [47, 180] loss: 1.564 [47, 240] loss: 1.559 [47, 300] loss: 1.554 [47, 360] loss: 1.561 Epoch: 47 -> Loss: 1.67219090462 Epoch: 47 -> Test Accuracy: 38.36 [48, 60] loss: 1.574 [48, 120] loss: 1.550 [48, 180] loss: 1.553 [48, 240] loss: 1.554 [48, 300] loss: 1.558 [48, 360] loss: 1.562 Epoch: 48 -> Loss: 1.62607038021 Epoch: 48 -> Test Accuracy: 38.48 [49, 60] loss: 1.541 [49, 120] loss: 1.557 [49, 180] loss: 1.554 [49, 240] loss: 1.560 [49, 300] loss: 1.560 [49, 360] loss: 1.567 Epoch: 49 -> Loss: 1.83787608147 Epoch: 49 -> Test Accuracy: 38.48 [50, 60] loss: 1.544 [50, 120] loss: 1.565 [50, 180] loss: 1.569 [50, 240] loss: 1.548 [50, 300] loss: 1.553 [50, 360] loss: 1.547 Epoch: 50 -> Loss: 1.45148348808 Epoch: 50 -> Test Accuracy: 38.48 [51, 60] loss: 1.569 [51, 120] loss: 1.558 [51, 180] loss: 1.549 [51, 240] loss: 1.546 [51, 300] loss: 1.535 [51, 360] loss: 1.531 Epoch: 51 -> Loss: 1.78328168392 Epoch: 51 -> Test Accuracy: 38.39 [52, 60] loss: 1.561 [52, 120] loss: 1.558 [52, 180] loss: 1.554 [52, 240] loss: 1.543 [52, 300] loss: 1.559 [52, 360] loss: 1.551 Epoch: 52 -> Loss: 1.56824326515 Epoch: 52 -> Test Accuracy: 38.17 [53, 60] loss: 1.556 [53, 120] loss: 1.569 [53, 180] loss: 1.543 [53, 240] loss: 1.555 [53, 300] loss: 1.557 [53, 360] loss: 1.560 Epoch: 53 -> Loss: 1.67066383362 Epoch: 53 -> Test Accuracy: 38.47 [54, 60] loss: 1.553 [54, 120] loss: 1.552 [54, 180] loss: 1.552 [54, 240] loss: 1.547 [54, 300] loss: 1.537 [54, 360] loss: 1.551 Epoch: 54 -> Loss: 1.50185477734 Epoch: 54 -> Test Accuracy: 38.33 [55, 60] loss: 1.561 [55, 120] loss: 1.526 [55, 180] loss: 1.549 [55, 240] loss: 1.554 [55, 300] loss: 1.539 [55, 360] loss: 1.548 Epoch: 55 -> Loss: 1.56229543686 Epoch: 55 -> Test Accuracy: 38.22 [56, 60] loss: 1.539 [56, 120] loss: 1.555 [56, 180] loss: 1.555 [56, 240] loss: 1.555 [56, 300] loss: 1.521 [56, 360] loss: 1.536 Epoch: 56 -> Loss: 1.52365350723 Epoch: 56 -> Test Accuracy: 38.33 [57, 60] loss: 1.546 [57, 120] loss: 1.544 [57, 180] loss: 1.536 [57, 240] loss: 1.555 [57, 300] loss: 1.545 [57, 360] loss: 1.534 Epoch: 57 -> Loss: 1.40506565571 Epoch: 57 -> Test Accuracy: 38.4 [58, 60] loss: 1.549 [58, 120] loss: 1.549 [58, 180] loss: 1.534 [58, 240] loss: 1.535 [58, 300] loss: 1.573 [58, 360] loss: 1.547 Epoch: 58 -> Loss: 1.54378581047 Epoch: 58 -> Test Accuracy: 38.48 [59, 60] loss: 1.550 [59, 120] loss: 1.531 [59, 180] loss: 1.543 [59, 240] loss: 1.548 [59, 300] loss: 1.546 [59, 360] loss: 1.544 Epoch: 59 -> Loss: 1.51378953457 Epoch: 59 -> Test Accuracy: 38.41 [60, 60] loss: 1.550 [60, 120] loss: 1.552 [60, 180] loss: 1.540 [60, 240] loss: 1.550 [60, 300] loss: 1.537 [60, 360] loss: 1.544 Epoch: 60 -> Loss: 1.68436467648 Epoch: 60 -> Test Accuracy: 38.28 [61, 60] loss: 1.536 [61, 120] loss: 1.549 [61, 180] loss: 1.535 [61, 240] loss: 1.544 [61, 300] loss: 1.546 [61, 360] loss: 1.542 Epoch: 61 -> Loss: 1.76267027855 Epoch: 61 -> Test Accuracy: 38.55 [62, 60] loss: 1.537 [62, 120] loss: 1.548 [62, 180] loss: 1.539 [62, 240] loss: 1.542 [62, 300] loss: 1.546 [62, 360] loss: 1.529 Epoch: 62 -> Loss: 1.42316508293 Epoch: 62 -> Test Accuracy: 38.47 [63, 60] loss: 1.543 [63, 120] loss: 1.547 [63, 180] loss: 1.526 [63, 240] loss: 1.539 [63, 300] loss: 1.526 [63, 360] loss: 1.545 Epoch: 63 -> Loss: 1.53315424919 Epoch: 63 -> Test Accuracy: 38.65 [64, 60] loss: 1.554 [64, 120] loss: 1.521 [64, 180] loss: 1.531 [64, 240] loss: 1.554 [64, 300] loss: 1.559 [64, 360] loss: 1.557 Epoch: 64 -> Loss: 1.57683038712 Epoch: 64 -> Test Accuracy: 38.31 [65, 60] loss: 1.536 [65, 120] loss: 1.544 [65, 180] loss: 1.519 [65, 240] loss: 1.539 [65, 300] loss: 1.543 [65, 360] loss: 1.538 Epoch: 65 -> Loss: 1.5305492878 Epoch: 65 -> Test Accuracy: 38.49 [66, 60] loss: 1.546 [66, 120] loss: 1.525 [66, 180] loss: 1.541 [66, 240] loss: 1.545 [66, 300] loss: 1.535 [66, 360] loss: 1.540 Epoch: 66 -> Loss: 1.58117127419 Epoch: 66 -> Test Accuracy: 38.48 [67, 60] loss: 1.538 [67, 120] loss: 1.539 [67, 180] loss: 1.542 [67, 240] loss: 1.540 [67, 300] loss: 1.548 [67, 360] loss: 1.549 Epoch: 67 -> Loss: 1.50874257088 Epoch: 67 -> Test Accuracy: 38.59 [68, 60] loss: 1.516 [68, 120] loss: 1.539 [68, 180] loss: 1.544 [68, 240] loss: 1.534 [68, 300] loss: 1.531 [68, 360] loss: 1.541 Epoch: 68 -> Loss: 1.54752898216 Epoch: 68 -> Test Accuracy: 38.49 [69, 60] loss: 1.510 [69, 120] loss: 1.540 [69, 180] loss: 1.557 [69, 240] loss: 1.539 [69, 300] loss: 1.539 [69, 360] loss: 1.555 Epoch: 69 -> Loss: 1.28347504139 Epoch: 69 -> Test Accuracy: 38.78 [70, 60] loss: 1.546 [70, 120] loss: 1.540 [70, 180] loss: 1.525 [70, 240] loss: 1.528 [70, 300] loss: 1.533 [70, 360] loss: 1.524 Epoch: 70 -> Loss: 1.40877139568 Epoch: 70 -> Test Accuracy: 38.35 [71, 60] loss: 1.513 [71, 120] loss: 1.542 [71, 180] loss: 1.527 [71, 240] loss: 1.541 [71, 300] loss: 1.515 [71, 360] loss: 1.541 Epoch: 71 -> Loss: 1.47087216377 Epoch: 71 -> Test Accuracy: 38.35 [72, 60] loss: 1.545 [72, 120] loss: 1.528 [72, 180] loss: 1.547 [72, 240] loss: 1.541 [72, 300] loss: 1.549 [72, 360] loss: 1.530 Epoch: 72 -> Loss: 1.65349137783 Epoch: 72 -> Test Accuracy: 38.49 [73, 60] loss: 1.528 [73, 120] loss: 1.543 [73, 180] loss: 1.539 [73, 240] loss: 1.529 [73, 300] loss: 1.541 [73, 360] loss: 1.520 Epoch: 73 -> Loss: 1.44064891338 Epoch: 73 -> Test Accuracy: 38.76 [74, 60] loss: 1.519 [74, 120] loss: 1.541 [74, 180] loss: 1.524 [74, 240] loss: 1.537 [74, 300] loss: 1.524 [74, 360] loss: 1.533 Epoch: 74 -> Loss: 1.42219233513 Epoch: 74 -> Test Accuracy: 38.7 [75, 60] loss: 1.536 [75, 120] loss: 1.516 [75, 180] loss: 1.549 [75, 240] loss: 1.517 [75, 300] loss: 1.539 [75, 360] loss: 1.526 Epoch: 75 -> Loss: 1.5182813406 Epoch: 75 -> Test Accuracy: 38.65 [76, 60] loss: 1.551 [76, 120] loss: 1.532 [76, 180] loss: 1.551 [76, 240] loss: 1.526 [76, 300] loss: 1.520 [76, 360] loss: 1.530 Epoch: 76 -> Loss: 1.56579899788 Epoch: 76 -> Test Accuracy: 38.66 [77, 60] loss: 1.524 [77, 120] loss: 1.531 [77, 180] loss: 1.531 [77, 240] loss: 1.533 [77, 300] loss: 1.512 [77, 360] loss: 1.549 Epoch: 77 -> Loss: 1.382553339 Epoch: 77 -> Test Accuracy: 38.44 [78, 60] loss: 1.544 [78, 120] loss: 1.522 [78, 180] loss: 1.522 [78, 240] loss: 1.536 [78, 300] loss: 1.520 [78, 360] loss: 1.518 Epoch: 78 -> Loss: 1.4872071743 Epoch: 78 -> Test Accuracy: 38.69 [79, 60] loss: 1.527 [79, 120] loss: 1.524 [79, 180] loss: 1.548 [79, 240] loss: 1.495 [79, 300] loss: 1.531 [79, 360] loss: 1.535 Epoch: 79 -> Loss: 1.6630461216 Epoch: 79 -> Test Accuracy: 38.66 [80, 60] loss: 1.524 [80, 120] loss: 1.539 [80, 180] loss: 1.524 [80, 240] loss: 1.521 [80, 300] loss: 1.527 [80, 360] loss: 1.518 Epoch: 80 -> Loss: 1.40660357475 Epoch: 80 -> Test Accuracy: 38.56 [81, 60] loss: 1.534 [81, 120] loss: 1.524 [81, 180] loss: 1.523 [81, 240] loss: 1.541 [81, 300] loss: 1.539 [81, 360] loss: 1.513 Epoch: 81 -> Loss: 1.63141596317 Epoch: 81 -> Test Accuracy: 38.78 [82, 60] loss: 1.529 [82, 120] loss: 1.526 [82, 180] loss: 1.536 [82, 240] loss: 1.525 [82, 300] loss: 1.516 [82, 360] loss: 1.537 Epoch: 82 -> Loss: 1.54906511307 Epoch: 82 -> Test Accuracy: 38.61 [83, 60] loss: 1.516 [83, 120] loss: 1.523 [83, 180] loss: 1.533 [83, 240] loss: 1.550 [83, 300] loss: 1.528 [83, 360] loss: 1.548 Epoch: 83 -> Loss: 1.86146092415 Epoch: 83 -> Test Accuracy: 38.57 [84, 60] loss: 1.512 [84, 120] loss: 1.529 [84, 180] loss: 1.522 [84, 240] loss: 1.526 [84, 300] loss: 1.525 [84, 360] loss: 1.535 Epoch: 84 -> Loss: 1.56655955315 Epoch: 84 -> Test Accuracy: 38.65 [85, 60] loss: 1.535 [85, 120] loss: 1.540 [85, 180] loss: 1.522 [85, 240] loss: 1.534 [85, 300] loss: 1.548 [85, 360] loss: 1.516 Epoch: 85 -> Loss: 1.41678702831 Epoch: 85 -> Test Accuracy: 38.65 [86, 60] loss: 1.523 [86, 120] loss: 1.532 [86, 180] loss: 1.542 [86, 240] loss: 1.520 [86, 300] loss: 1.523 [86, 360] loss: 1.529 Epoch: 86 -> Loss: 1.42956662178 Epoch: 86 -> Test Accuracy: 38.94 [87, 60] loss: 1.527 [87, 120] loss: 1.527 [87, 180] loss: 1.542 [87, 240] loss: 1.530 [87, 300] loss: 1.509 [87, 360] loss: 1.536 Epoch: 87 -> Loss: 1.61188793182 Epoch: 87 -> Test Accuracy: 38.77 [88, 60] loss: 1.530 [88, 120] loss: 1.529 [88, 180] loss: 1.520 [88, 240] loss: 1.541 [88, 300] loss: 1.538 [88, 360] loss: 1.522 Epoch: 88 -> Loss: 1.42896151543 Epoch: 88 -> Test Accuracy: 38.6 [89, 60] loss: 1.528 [89, 120] loss: 1.526 [89, 180] loss: 1.549 [89, 240] loss: 1.510 [89, 300] loss: 1.519 [89, 360] loss: 1.518 Epoch: 89 -> Loss: 1.43834519386 Epoch: 89 -> Test Accuracy: 38.86 [90, 60] loss: 1.519 [90, 120] loss: 1.515 [90, 180] loss: 1.520 [90, 240] loss: 1.523 [90, 300] loss: 1.504 [90, 360] loss: 1.527 Epoch: 90 -> Loss: 1.6965328455 Epoch: 90 -> Test Accuracy: 38.68 [91, 60] loss: 1.523 [91, 120] loss: 1.529 [91, 180] loss: 1.537 [91, 240] loss: 1.526 [91, 300] loss: 1.534 [91, 360] loss: 1.510 Epoch: 91 -> Loss: 1.49568963051 Epoch: 91 -> Test Accuracy: 38.68 [92, 60] loss: 1.534 [92, 120] loss: 1.534 [92, 180] loss: 1.536 [92, 240] loss: 1.513 [92, 300] loss: 1.522 [92, 360] loss: 1.535 Epoch: 92 -> Loss: 1.63094496727 Epoch: 92 -> Test Accuracy: 38.84 [93, 60] loss: 1.528 [93, 120] loss: 1.522 [93, 180] loss: 1.515 [93, 240] loss: 1.535 [93, 300] loss: 1.524 [93, 360] loss: 1.529 Epoch: 93 -> Loss: 1.35829985142 Epoch: 93 -> Test Accuracy: 38.59 [94, 60] loss: 1.528 [94, 120] loss: 1.512 [94, 180] loss: 1.518 [94, 240] loss: 1.519 [94, 300] loss: 1.543 [94, 360] loss: 1.532 Epoch: 94 -> Loss: 1.53057134151 Epoch: 94 -> Test Accuracy: 38.54 [95, 60] loss: 1.519 [95, 120] loss: 1.532 [95, 180] loss: 1.519 [95, 240] loss: 1.538 [95, 300] loss: 1.525 [95, 360] loss: 1.513 Epoch: 95 -> Loss: 1.56014883518 Epoch: 95 -> Test Accuracy: 38.93 [96, 60] loss: 1.526 [96, 120] loss: 1.506 [96, 180] loss: 1.518 [96, 240] loss: 1.514 [96, 300] loss: 1.505 [96, 360] loss: 1.518 Epoch: 96 -> Loss: 1.57741713524 Epoch: 96 -> Test Accuracy: 38.83 [97, 60] loss: 1.518 [97, 120] loss: 1.513 [97, 180] loss: 1.537 [97, 240] loss: 1.508 [97, 300] loss: 1.525 [97, 360] loss: 1.532 Epoch: 97 -> Loss: 1.39256501198 Epoch: 97 -> Test Accuracy: 39.0 [98, 60] loss: 1.509 [98, 120] loss: 1.503 [98, 180] loss: 1.532 [98, 240] loss: 1.503 [98, 300] loss: 1.518 [98, 360] loss: 1.534 Epoch: 98 -> Loss: 1.58949923515 Epoch: 98 -> Test Accuracy: 38.64 [99, 60] loss: 1.506 [99, 120] loss: 1.527 [99, 180] loss: 1.523 [99, 240] loss: 1.520 [99, 300] loss: 1.511 [99, 360] loss: 1.518 Epoch: 99 -> Loss: 1.3212211132 Epoch: 99 -> Test Accuracy: 38.87 [100, 60] loss: 1.531 [100, 120] loss: 1.494 [100, 180] loss: 1.527 [100, 240] loss: 1.519 [100, 300] loss: 1.531 [100, 360] loss: 1.524 Epoch: 100 -> Loss: 1.61050927639 Epoch: 100 -> Test Accuracy: 39.12 Finished Training
In [9]:
# train ConvClassifiers on feature map of net_3block
conv_block5_loss_log, _, conv_block5_test_accuracy_log, _, _ = tr.train_all_blocks(5, 10, [0.1, 0.02, 0.004, 0.0008],
[35, 70, 85, 100], 0.9, 5e-4, net_block5, criterion, trainloader, None, testloader, use_ConvClassifier=True)
[1, 60] loss: 1.352 [1, 120] loss: 1.061 [1, 180] loss: 0.948 [1, 240] loss: 0.928 [1, 300] loss: 0.862 [1, 360] loss: 0.831 Epoch: 1 -> Loss: 0.759939789772 Epoch: 1 -> Test Accuracy: 68.27 [2, 60] loss: 0.776 [2, 120] loss: 0.762 [2, 180] loss: 0.733 [2, 240] loss: 0.739 [2, 300] loss: 0.712 [2, 360] loss: 0.707 Epoch: 2 -> Loss: 0.522010505199 Epoch: 2 -> Test Accuracy: 74.42 [3, 60] loss: 0.652 [3, 120] loss: 0.656 [3, 180] loss: 0.670 [3, 240] loss: 0.644 [3, 300] loss: 0.668 [3, 360] loss: 0.649 Epoch: 3 -> Loss: 0.641436398029 Epoch: 3 -> Test Accuracy: 75.58 [4, 60] loss: 0.594 [4, 120] loss: 0.609 [4, 180] loss: 0.618 [4, 240] loss: 0.615 [4, 300] loss: 0.610 [4, 360] loss: 0.612 Epoch: 4 -> Loss: 0.61285841465 Epoch: 4 -> Test Accuracy: 75.85 [5, 60] loss: 0.570 [5, 120] loss: 0.588 [5, 180] loss: 0.560 [5, 240] loss: 0.563 [5, 300] loss: 0.579 [5, 360] loss: 0.582 Epoch: 5 -> Loss: 0.694158673286 Epoch: 5 -> Test Accuracy: 75.9 [6, 60] loss: 0.559 [6, 120] loss: 0.558 [6, 180] loss: 0.545 [6, 240] loss: 0.585 [6, 300] loss: 0.541 [6, 360] loss: 0.553 Epoch: 6 -> Loss: 0.620222210884 Epoch: 6 -> Test Accuracy: 78.7 [7, 60] loss: 0.528 [7, 120] loss: 0.532 [7, 180] loss: 0.532 [7, 240] loss: 0.534 [7, 300] loss: 0.535 [7, 360] loss: 0.544 Epoch: 7 -> Loss: 0.471471130848 Epoch: 7 -> Test Accuracy: 78.89 [8, 60] loss: 0.518 [8, 120] loss: 0.506 [8, 180] loss: 0.523 [8, 240] loss: 0.527 [8, 300] loss: 0.542 [8, 360] loss: 0.521 Epoch: 8 -> Loss: 0.526454925537 Epoch: 8 -> Test Accuracy: 79.92 [9, 60] loss: 0.491 [9, 120] loss: 0.508 [9, 180] loss: 0.511 [9, 240] loss: 0.520 [9, 300] loss: 0.513 [9, 360] loss: 0.520 Epoch: 9 -> Loss: 0.396468639374 Epoch: 9 -> Test Accuracy: 77.85 [10, 60] loss: 0.481 [10, 120] loss: 0.495 [10, 180] loss: 0.509 [10, 240] loss: 0.508 [10, 300] loss: 0.509 [10, 360] loss: 0.490 Epoch: 10 -> Loss: 0.374304443598 Epoch: 10 -> Test Accuracy: 77.9 [11, 60] loss: 0.478 [11, 120] loss: 0.501 [11, 180] loss: 0.484 [11, 240] loss: 0.510 [11, 300] loss: 0.499 [11, 360] loss: 0.491 Epoch: 11 -> Loss: 0.547458052635 Epoch: 11 -> Test Accuracy: 79.02 [12, 60] loss: 0.471 [12, 120] loss: 0.486 [12, 180] loss: 0.496 [12, 240] loss: 0.475 [12, 300] loss: 0.504 [12, 360] loss: 0.497 Epoch: 12 -> Loss: 0.513858675957 Epoch: 12 -> Test Accuracy: 79.76 [13, 60] loss: 0.451 [13, 120] loss: 0.459 [13, 180] loss: 0.501 [13, 240] loss: 0.472 [13, 300] loss: 0.472 [13, 360] loss: 0.495 Epoch: 13 -> Loss: 0.468185126781 Epoch: 13 -> Test Accuracy: 78.25 [14, 60] loss: 0.465 [14, 120] loss: 0.480 [14, 180] loss: 0.462 [14, 240] loss: 0.471 [14, 300] loss: 0.478 [14, 360] loss: 0.485 Epoch: 14 -> Loss: 0.514837741852 Epoch: 14 -> Test Accuracy: 78.66 [15, 60] loss: 0.450 [15, 120] loss: 0.475 [15, 180] loss: 0.466 [15, 240] loss: 0.475 [15, 300] loss: 0.472 [15, 360] loss: 0.479 Epoch: 15 -> Loss: 0.471237182617 Epoch: 15 -> Test Accuracy: 80.28 [16, 60] loss: 0.446 [16, 120] loss: 0.446 [16, 180] loss: 0.456 [16, 240] loss: 0.481 [16, 300] loss: 0.480 [16, 360] loss: 0.489 Epoch: 16 -> Loss: 0.407373130322 Epoch: 16 -> Test Accuracy: 80.24 [17, 60] loss: 0.437 [17, 120] loss: 0.460 [17, 180] loss: 0.472 [17, 240] loss: 0.472 [17, 300] loss: 0.473 [17, 360] loss: 0.459 Epoch: 17 -> Loss: 0.510019779205 Epoch: 17 -> Test Accuracy: 81.08 [18, 60] loss: 0.438 [18, 120] loss: 0.441 [18, 180] loss: 0.471 [18, 240] loss: 0.481 [18, 300] loss: 0.467 [18, 360] loss: 0.464 Epoch: 18 -> Loss: 0.474419116974 Epoch: 18 -> Test Accuracy: 79.9 [19, 60] loss: 0.439 [19, 120] loss: 0.460 [19, 180] loss: 0.447 [19, 240] loss: 0.460 [19, 300] loss: 0.464 [19, 360] loss: 0.460 Epoch: 19 -> Loss: 0.493436008692 Epoch: 19 -> Test Accuracy: 80.47 [20, 60] loss: 0.428 [20, 120] loss: 0.444 [20, 180] loss: 0.455 [20, 240] loss: 0.460 [20, 300] loss: 0.456 [20, 360] loss: 0.476 Epoch: 20 -> Loss: 0.458377212286 Epoch: 20 -> Test Accuracy: 79.46 [21, 60] loss: 0.446 [21, 120] loss: 0.442 [21, 180] loss: 0.453 [21, 240] loss: 0.450 [21, 300] loss: 0.450 [21, 360] loss: 0.442 Epoch: 21 -> Loss: 0.395490467548 Epoch: 21 -> Test Accuracy: 79.5 [22, 60] loss: 0.417 [22, 120] loss: 0.430 [22, 180] loss: 0.466 [22, 240] loss: 0.441 [22, 300] loss: 0.462 [22, 360] loss: 0.479 Epoch: 22 -> Loss: 0.603022575378 Epoch: 22 -> Test Accuracy: 81.12 [23, 60] loss: 0.436 [23, 120] loss: 0.420 [23, 180] loss: 0.430 [23, 240] loss: 0.457 [23, 300] loss: 0.456 [23, 360] loss: 0.460 Epoch: 23 -> Loss: 0.583153009415 Epoch: 23 -> Test Accuracy: 80.43 [24, 60] loss: 0.432 [24, 120] loss: 0.451 [24, 180] loss: 0.429 [24, 240] loss: 0.467 [24, 300] loss: 0.445 [24, 360] loss: 0.461 Epoch: 24 -> Loss: 0.404473781586 Epoch: 24 -> Test Accuracy: 80.75 [25, 60] loss: 0.420 [25, 120] loss: 0.424 [25, 180] loss: 0.449 [25, 240] loss: 0.457 [25, 300] loss: 0.448 [25, 360] loss: 0.464 Epoch: 25 -> Loss: 0.699390470982 Epoch: 25 -> Test Accuracy: 79.69 [26, 60] loss: 0.424 [26, 120] loss: 0.433 [26, 180] loss: 0.450 [26, 240] loss: 0.445 [26, 300] loss: 0.446 [26, 360] loss: 0.468 Epoch: 26 -> Loss: 0.464469283819 Epoch: 26 -> Test Accuracy: 80.1 [27, 60] loss: 0.433 [27, 120] loss: 0.423 [27, 180] loss: 0.420 [27, 240] loss: 0.451 [27, 300] loss: 0.445 [27, 360] loss: 0.460 Epoch: 27 -> Loss: 0.251999527216 Epoch: 27 -> Test Accuracy: 80.29 [28, 60] loss: 0.418 [28, 120] loss: 0.426 [28, 180] loss: 0.438 [28, 240] loss: 0.441 [28, 300] loss: 0.442 [28, 360] loss: 0.460 Epoch: 28 -> Loss: 0.329118907452 Epoch: 28 -> Test Accuracy: 81.82 [29, 60] loss: 0.411 [29, 120] loss: 0.421 [29, 180] loss: 0.445 [29, 240] loss: 0.423 [29, 300] loss: 0.450 [29, 360] loss: 0.466 Epoch: 29 -> Loss: 0.509974241257 Epoch: 29 -> Test Accuracy: 79.72 [30, 60] loss: 0.401 [30, 120] loss: 0.413 [30, 180] loss: 0.446 [30, 240] loss: 0.453 [30, 300] loss: 0.464 [30, 360] loss: 0.441 Epoch: 30 -> Loss: 0.528644680977 Epoch: 30 -> Test Accuracy: 81.07 [31, 60] loss: 0.411 [31, 120] loss: 0.429 [31, 180] loss: 0.420 [31, 240] loss: 0.447 [31, 300] loss: 0.464 [31, 360] loss: 0.473 Epoch: 31 -> Loss: 0.543130278587 Epoch: 31 -> Test Accuracy: 80.92 [32, 60] loss: 0.438 [32, 120] loss: 0.416 [32, 180] loss: 0.414 [32, 240] loss: 0.469 [32, 300] loss: 0.440 [32, 360] loss: 0.436 Epoch: 32 -> Loss: 0.521073997021 Epoch: 32 -> Test Accuracy: 80.44 [33, 60] loss: 0.431 [33, 120] loss: 0.417 [33, 180] loss: 0.445 [33, 240] loss: 0.429 [33, 300] loss: 0.454 [33, 360] loss: 0.450 Epoch: 33 -> Loss: 0.466006666422 Epoch: 33 -> Test Accuracy: 80.63 [34, 60] loss: 0.403 [34, 120] loss: 0.421 [34, 180] loss: 0.444 [34, 240] loss: 0.445 [34, 300] loss: 0.442 [34, 360] loss: 0.441 Epoch: 34 -> Loss: 0.361333429813 Epoch: 34 -> Test Accuracy: 80.83 [35, 60] loss: 0.396 [35, 120] loss: 0.424 [35, 180] loss: 0.431 [35, 240] loss: 0.427 [35, 300] loss: 0.432 [35, 360] loss: 0.439 Epoch: 35 -> Loss: 0.508682787418 Epoch: 35 -> Test Accuracy: 81.01 [36, 60] loss: 0.347 [36, 120] loss: 0.314 [36, 180] loss: 0.307 [36, 240] loss: 0.306 [36, 300] loss: 0.311 [36, 360] loss: 0.296 Epoch: 36 -> Loss: 0.330245882273 Epoch: 36 -> Test Accuracy: 84.61 [37, 60] loss: 0.283 [37, 120] loss: 0.278 [37, 180] loss: 0.277 [37, 240] loss: 0.269 [37, 300] loss: 0.279 [37, 360] loss: 0.280 Epoch: 37 -> Loss: 0.155964404345 Epoch: 37 -> Test Accuracy: 84.76 [38, 60] loss: 0.255 [38, 120] loss: 0.263 [38, 180] loss: 0.258 [38, 240] loss: 0.267 [38, 300] loss: 0.263 [38, 360] loss: 0.272 Epoch: 38 -> Loss: 0.415426552296 Epoch: 38 -> Test Accuracy: 85.03 [39, 60] loss: 0.247 [39, 120] loss: 0.247 [39, 180] loss: 0.257 [39, 240] loss: 0.259 [39, 300] loss: 0.256 [39, 360] loss: 0.251 Epoch: 39 -> Loss: 0.140538573265 Epoch: 39 -> Test Accuracy: 84.63 [40, 60] loss: 0.238 [40, 120] loss: 0.247 [40, 180] loss: 0.251 [40, 240] loss: 0.250 [40, 300] loss: 0.251 [40, 360] loss: 0.271 Epoch: 40 -> Loss: 0.245443612337 Epoch: 40 -> Test Accuracy: 84.15 [41, 60] loss: 0.232 [41, 120] loss: 0.232 [41, 180] loss: 0.249 [41, 240] loss: 0.250 [41, 300] loss: 0.259 [41, 360] loss: 0.261 Epoch: 41 -> Loss: 0.186827391386 Epoch: 41 -> Test Accuracy: 84.33 [42, 60] loss: 0.237 [42, 120] loss: 0.235 [42, 180] loss: 0.246 [42, 240] loss: 0.240 [42, 300] loss: 0.250 [42, 360] loss: 0.255 Epoch: 42 -> Loss: 0.281368166208 Epoch: 42 -> Test Accuracy: 84.08 [43, 60] loss: 0.235 [43, 120] loss: 0.248 [43, 180] loss: 0.240 [43, 240] loss: 0.253 [43, 300] loss: 0.245 [43, 360] loss: 0.251 Epoch: 43 -> Loss: 0.160958588123 Epoch: 43 -> Test Accuracy: 84.44 [44, 60] loss: 0.233 [44, 120] loss: 0.239 [44, 180] loss: 0.238 [44, 240] loss: 0.247 [44, 300] loss: 0.253 [44, 360] loss: 0.244 Epoch: 44 -> Loss: 0.334113538265 Epoch: 44 -> Test Accuracy: 83.81 [45, 60] loss: 0.227 [45, 120] loss: 0.227 [45, 180] loss: 0.242 [45, 240] loss: 0.259 [45, 300] loss: 0.248 [45, 360] loss: 0.240 Epoch: 45 -> Loss: 0.149308085442 Epoch: 45 -> Test Accuracy: 84.51 [46, 60] loss: 0.228 [46, 120] loss: 0.234 [46, 180] loss: 0.245 [46, 240] loss: 0.243 [46, 300] loss: 0.241 [46, 360] loss: 0.257 Epoch: 46 -> Loss: 0.253108352423 Epoch: 46 -> Test Accuracy: 83.79 [47, 60] loss: 0.224 [47, 120] loss: 0.233 [47, 180] loss: 0.245 [47, 240] loss: 0.252 [47, 300] loss: 0.253 [47, 360] loss: 0.255 Epoch: 47 -> Loss: 0.322635650635 Epoch: 47 -> Test Accuracy: 83.32 [48, 60] loss: 0.237 [48, 120] loss: 0.229 [48, 180] loss: 0.242 [48, 240] loss: 0.247 [48, 300] loss: 0.250 [48, 360] loss: 0.256 Epoch: 48 -> Loss: 0.182466894388 Epoch: 48 -> Test Accuracy: 84.26 [49, 60] loss: 0.232 [49, 120] loss: 0.235 [49, 180] loss: 0.232 [49, 240] loss: 0.246 [49, 300] loss: 0.249 [49, 360] loss: 0.251 Epoch: 49 -> Loss: 0.236679837108 Epoch: 49 -> Test Accuracy: 84.2 [50, 60] loss: 0.223 [50, 120] loss: 0.230 [50, 180] loss: 0.230 [50, 240] loss: 0.256 [50, 300] loss: 0.250 [50, 360] loss: 0.252 Epoch: 50 -> Loss: 0.367073625326 Epoch: 50 -> Test Accuracy: 83.6 [51, 60] loss: 0.222 [51, 120] loss: 0.233 [51, 180] loss: 0.244 [51, 240] loss: 0.232 [51, 300] loss: 0.261 [51, 360] loss: 0.252 Epoch: 51 -> Loss: 0.276985824108 Epoch: 51 -> Test Accuracy: 84.11 [52, 60] loss: 0.226 [52, 120] loss: 0.222 [52, 180] loss: 0.231 [52, 240] loss: 0.239 [52, 300] loss: 0.264 [52, 360] loss: 0.271 Epoch: 52 -> Loss: 0.399486720562 Epoch: 52 -> Test Accuracy: 84.36 [53, 60] loss: 0.222 [53, 120] loss: 0.225 [53, 180] loss: 0.237 [53, 240] loss: 0.247 [53, 300] loss: 0.242 [53, 360] loss: 0.264 Epoch: 53 -> Loss: 0.184724837542 Epoch: 53 -> Test Accuracy: 83.78 [54, 60] loss: 0.218 [54, 120] loss: 0.227 [54, 180] loss: 0.249 [54, 240] loss: 0.238 [54, 300] loss: 0.247 [54, 360] loss: 0.256 Epoch: 54 -> Loss: 0.230543702841 Epoch: 54 -> Test Accuracy: 83.59 [55, 60] loss: 0.219 [55, 120] loss: 0.232 [55, 180] loss: 0.250 [55, 240] loss: 0.236 [55, 300] loss: 0.246 [55, 360] loss: 0.246 Epoch: 55 -> Loss: 0.183805555105 Epoch: 55 -> Test Accuracy: 84.3 [56, 60] loss: 0.226 [56, 120] loss: 0.248 [56, 180] loss: 0.251 [56, 240] loss: 0.238 [56, 300] loss: 0.234 [56, 360] loss: 0.246 Epoch: 56 -> Loss: 0.223450392485 Epoch: 56 -> Test Accuracy: 84.09 [57, 60] loss: 0.219 [57, 120] loss: 0.221 [57, 180] loss: 0.235 [57, 240] loss: 0.237 [57, 300] loss: 0.253 [57, 360] loss: 0.253 Epoch: 57 -> Loss: 0.286977887154 Epoch: 57 -> Test Accuracy: 83.37 [58, 60] loss: 0.222 [58, 120] loss: 0.231 [58, 180] loss: 0.228 [58, 240] loss: 0.228 [58, 300] loss: 0.235 [58, 360] loss: 0.248 Epoch: 58 -> Loss: 0.26318192482 Epoch: 58 -> Test Accuracy: 83.87 [59, 60] loss: 0.218 [59, 120] loss: 0.238 [59, 180] loss: 0.242 [59, 240] loss: 0.221 [59, 300] loss: 0.243 [59, 360] loss: 0.240 Epoch: 59 -> Loss: 0.301249235868 Epoch: 59 -> Test Accuracy: 83.58 [60, 60] loss: 0.229 [60, 120] loss: 0.223 [60, 180] loss: 0.228 [60, 240] loss: 0.251 [60, 300] loss: 0.242 [60, 360] loss: 0.249 Epoch: 60 -> Loss: 0.272841185331 Epoch: 60 -> Test Accuracy: 83.42 [61, 60] loss: 0.217 [61, 120] loss: 0.230 [61, 180] loss: 0.231 [61, 240] loss: 0.245 [61, 300] loss: 0.245 [61, 360] loss: 0.248 Epoch: 61 -> Loss: 0.337931156158 Epoch: 61 -> Test Accuracy: 83.79 [62, 60] loss: 0.221 [62, 120] loss: 0.224 [62, 180] loss: 0.234 [62, 240] loss: 0.235 [62, 300] loss: 0.246 [62, 360] loss: 0.240 Epoch: 62 -> Loss: 0.36541262269 Epoch: 62 -> Test Accuracy: 83.66 [63, 60] loss: 0.218 [63, 120] loss: 0.225 [63, 180] loss: 0.248 [63, 240] loss: 0.229 [63, 300] loss: 0.241 [63, 360] loss: 0.250 Epoch: 63 -> Loss: 0.294481903315 Epoch: 63 -> Test Accuracy: 83.92 [64, 60] loss: 0.202 [64, 120] loss: 0.228 [64, 180] loss: 0.235 [64, 240] loss: 0.232 [64, 300] loss: 0.238 [64, 360] loss: 0.246 Epoch: 64 -> Loss: 0.326318830252 Epoch: 64 -> Test Accuracy: 83.68 [65, 60] loss: 0.204 [65, 120] loss: 0.223 [65, 180] loss: 0.238 [65, 240] loss: 0.249 [65, 300] loss: 0.254 [65, 360] loss: 0.240 Epoch: 65 -> Loss: 0.297684848309 Epoch: 65 -> Test Accuracy: 83.3 [66, 60] loss: 0.225 [66, 120] loss: 0.213 [66, 180] loss: 0.234 [66, 240] loss: 0.230 [66, 300] loss: 0.242 [66, 360] loss: 0.235 Epoch: 66 -> Loss: 0.362421035767 Epoch: 66 -> Test Accuracy: 83.9 [67, 60] loss: 0.220 [67, 120] loss: 0.238 [67, 180] loss: 0.236 [67, 240] loss: 0.232 [67, 300] loss: 0.247 [67, 360] loss: 0.220 Epoch: 67 -> Loss: 0.27500808239 Epoch: 67 -> Test Accuracy: 83.92 [68, 60] loss: 0.216 [68, 120] loss: 0.218 [68, 180] loss: 0.230 [68, 240] loss: 0.237 [68, 300] loss: 0.242 [68, 360] loss: 0.246 Epoch: 68 -> Loss: 0.251587361097 Epoch: 68 -> Test Accuracy: 83.74 [69, 60] loss: 0.212 [69, 120] loss: 0.228 [69, 180] loss: 0.226 [69, 240] loss: 0.239 [69, 300] loss: 0.237 [69, 360] loss: 0.238 Epoch: 69 -> Loss: 0.279275119305 Epoch: 69 -> Test Accuracy: 83.69 [70, 60] loss: 0.207 [70, 120] loss: 0.222 [70, 180] loss: 0.245 [70, 240] loss: 0.233 [70, 300] loss: 0.233 [70, 360] loss: 0.239 Epoch: 70 -> Loss: 0.242570310831 Epoch: 70 -> Test Accuracy: 83.99 [71, 60] loss: 0.177 [71, 120] loss: 0.160 [71, 180] loss: 0.162 [71, 240] loss: 0.157 [71, 300] loss: 0.159 [71, 360] loss: 0.147 Epoch: 71 -> Loss: 0.249945372343 Epoch: 71 -> Test Accuracy: 85.75 [72, 60] loss: 0.134 [72, 120] loss: 0.136 [72, 180] loss: 0.140 [72, 240] loss: 0.144 [72, 300] loss: 0.142 [72, 360] loss: 0.152 Epoch: 72 -> Loss: 0.185048639774 Epoch: 72 -> Test Accuracy: 85.67 [73, 60] loss: 0.137 [73, 120] loss: 0.128 [73, 180] loss: 0.135 [73, 240] loss: 0.134 [73, 300] loss: 0.128 [73, 360] loss: 0.146 Epoch: 73 -> Loss: 0.11365352571 Epoch: 73 -> Test Accuracy: 85.99 [74, 60] loss: 0.129 [74, 120] loss: 0.128 [74, 180] loss: 0.129 [74, 240] loss: 0.136 [74, 300] loss: 0.129 [74, 360] loss: 0.141 Epoch: 74 -> Loss: 0.118794694543 Epoch: 74 -> Test Accuracy: 85.53 [75, 60] loss: 0.119 [75, 120] loss: 0.130 [75, 180] loss: 0.129 [75, 240] loss: 0.127 [75, 300] loss: 0.125 [75, 360] loss: 0.131 Epoch: 75 -> Loss: 0.110058307648 Epoch: 75 -> Test Accuracy: 85.65 [76, 60] loss: 0.120 [76, 120] loss: 0.125 [76, 180] loss: 0.123 [76, 240] loss: 0.120 [76, 300] loss: 0.131 [76, 360] loss: 0.127 Epoch: 76 -> Loss: 0.138767778873 Epoch: 76 -> Test Accuracy: 85.72 [77, 60] loss: 0.122 [77, 120] loss: 0.118 [77, 180] loss: 0.119 [77, 240] loss: 0.122 [77, 300] loss: 0.132 [77, 360] loss: 0.131 Epoch: 77 -> Loss: 0.105447307229 Epoch: 77 -> Test Accuracy: 85.36 [78, 60] loss: 0.115 [78, 120] loss: 0.118 [78, 180] loss: 0.116 [78, 240] loss: 0.127 [78, 300] loss: 0.122 [78, 360] loss: 0.123 Epoch: 78 -> Loss: 0.103502966464 Epoch: 78 -> Test Accuracy: 85.48 [79, 60] loss: 0.115 [79, 120] loss: 0.113 [79, 180] loss: 0.114 [79, 240] loss: 0.116 [79, 300] loss: 0.125 [79, 360] loss: 0.121 Epoch: 79 -> Loss: 0.10041461885 Epoch: 79 -> Test Accuracy: 85.47 [80, 60] loss: 0.110 [80, 120] loss: 0.118 [80, 180] loss: 0.115 [80, 240] loss: 0.115 [80, 300] loss: 0.114 [80, 360] loss: 0.119 Epoch: 80 -> Loss: 0.0979191586375 Epoch: 80 -> Test Accuracy: 85.61 [81, 60] loss: 0.110 [81, 120] loss: 0.112 [81, 180] loss: 0.111 [81, 240] loss: 0.117 [81, 300] loss: 0.120 [81, 360] loss: 0.113 Epoch: 81 -> Loss: 0.0923889875412 Epoch: 81 -> Test Accuracy: 85.11 [82, 60] loss: 0.110 [82, 120] loss: 0.109 [82, 180] loss: 0.109 [82, 240] loss: 0.118 [82, 300] loss: 0.114 [82, 360] loss: 0.114 Epoch: 82 -> Loss: 0.19514246285 Epoch: 82 -> Test Accuracy: 85.32 [83, 60] loss: 0.107 [83, 120] loss: 0.109 [83, 180] loss: 0.111 [83, 240] loss: 0.107 [83, 300] loss: 0.122 [83, 360] loss: 0.114 Epoch: 83 -> Loss: 0.0596535205841 Epoch: 83 -> Test Accuracy: 85.24 [84, 60] loss: 0.106 [84, 120] loss: 0.107 [84, 180] loss: 0.107 [84, 240] loss: 0.115 [84, 300] loss: 0.115 [84, 360] loss: 0.119 Epoch: 84 -> Loss: 0.050937525928 Epoch: 84 -> Test Accuracy: 85.17 [85, 60] loss: 0.108 [85, 120] loss: 0.114 [85, 180] loss: 0.112 [85, 240] loss: 0.113 [85, 300] loss: 0.111 [85, 360] loss: 0.115 Epoch: 85 -> Loss: 0.100176349282 Epoch: 85 -> Test Accuracy: 84.99 [86, 60] loss: 0.101 [86, 120] loss: 0.096 [86, 180] loss: 0.091 [86, 240] loss: 0.096 [86, 300] loss: 0.099 [86, 360] loss: 0.101 Epoch: 86 -> Loss: 0.0698599889874 Epoch: 86 -> Test Accuracy: 85.47 [87, 60] loss: 0.094 [87, 120] loss: 0.092 [87, 180] loss: 0.096 [87, 240] loss: 0.100 [87, 300] loss: 0.092 [87, 360] loss: 0.096 Epoch: 87 -> Loss: 0.139446765184 Epoch: 87 -> Test Accuracy: 85.51 [88, 60] loss: 0.096 [88, 120] loss: 0.093 [88, 180] loss: 0.094 [88, 240] loss: 0.093 [88, 300] loss: 0.095 [88, 360] loss: 0.094 Epoch: 88 -> Loss: 0.0789198949933 Epoch: 88 -> Test Accuracy: 85.44 [89, 60] loss: 0.097 [89, 120] loss: 0.091 [89, 180] loss: 0.091 [89, 240] loss: 0.090 [89, 300] loss: 0.098 [89, 360] loss: 0.091 Epoch: 89 -> Loss: 0.14526823163 Epoch: 89 -> Test Accuracy: 85.68 [90, 60] loss: 0.088 [90, 120] loss: 0.091 [90, 180] loss: 0.089 [90, 240] loss: 0.095 [90, 300] loss: 0.095 [90, 360] loss: 0.094 Epoch: 90 -> Loss: 0.108555816114 Epoch: 90 -> Test Accuracy: 85.54 [91, 60] loss: 0.092 [91, 120] loss: 0.090 [91, 180] loss: 0.089 [91, 240] loss: 0.091 [91, 300] loss: 0.090 [91, 360] loss: 0.098 Epoch: 91 -> Loss: 0.0691407769918 Epoch: 91 -> Test Accuracy: 85.56 [92, 60] loss: 0.089 [92, 120] loss: 0.088 [92, 180] loss: 0.090 [92, 240] loss: 0.097 [92, 300] loss: 0.089 [92, 360] loss: 0.088 Epoch: 92 -> Loss: 0.112366870046 Epoch: 92 -> Test Accuracy: 85.45 [93, 60] loss: 0.091 [93, 120] loss: 0.089 [93, 180] loss: 0.086 [93, 240] loss: 0.088 [93, 300] loss: 0.094 [93, 360] loss: 0.093 Epoch: 93 -> Loss: 0.0774373561144 Epoch: 93 -> Test Accuracy: 85.54 [94, 60] loss: 0.087 [94, 120] loss: 0.092 [94, 180] loss: 0.085 [94, 240] loss: 0.088 [94, 300] loss: 0.085 [94, 360] loss: 0.091 Epoch: 94 -> Loss: 0.156728237867 Epoch: 94 -> Test Accuracy: 85.41 [95, 60] loss: 0.085 [95, 120] loss: 0.086 [95, 180] loss: 0.091 [95, 240] loss: 0.090 [95, 300] loss: 0.088 [95, 360] loss: 0.090 Epoch: 95 -> Loss: 0.162043347955 Epoch: 95 -> Test Accuracy: 85.32 [96, 60] loss: 0.087 [96, 120] loss: 0.087 [96, 180] loss: 0.088 [96, 240] loss: 0.089 [96, 300] loss: 0.089 [96, 360] loss: 0.093 Epoch: 96 -> Loss: 0.0902237519622 Epoch: 96 -> Test Accuracy: 85.45 [97, 60] loss: 0.090 [97, 120] loss: 0.091 [97, 180] loss: 0.085 [97, 240] loss: 0.086 [97, 300] loss: 0.087 [97, 360] loss: 0.088 Epoch: 97 -> Loss: 0.119821645319 Epoch: 97 -> Test Accuracy: 85.47 [98, 60] loss: 0.086 [98, 120] loss: 0.085 [98, 180] loss: 0.086 [98, 240] loss: 0.087 [98, 300] loss: 0.090 [98, 360] loss: 0.094 Epoch: 98 -> Loss: 0.072603456676 Epoch: 98 -> Test Accuracy: 85.49 [99, 60] loss: 0.089 [99, 120] loss: 0.088 [99, 180] loss: 0.090 [99, 240] loss: 0.090 [99, 300] loss: 0.084 [99, 360] loss: 0.091 Epoch: 99 -> Loss: 0.125592559576 Epoch: 99 -> Test Accuracy: 85.5 [100, 60] loss: 0.082 [100, 120] loss: 0.085 [100, 180] loss: 0.092 [100, 240] loss: 0.089 [100, 300] loss: 0.083 [100, 360] loss: 0.090 Epoch: 100 -> Loss: 0.0403878800571 Epoch: 100 -> Test Accuracy: 85.28 Finished Training [1, 60] loss: 0.950 [1, 120] loss: 0.655 [1, 180] loss: 0.596 [1, 240] loss: 0.551 [1, 300] loss: 0.519 [1, 360] loss: 0.509 Epoch: 1 -> Loss: 0.539218306541 Epoch: 1 -> Test Accuracy: 79.69 [2, 60] loss: 0.457 [2, 120] loss: 0.450 [2, 180] loss: 0.449 [2, 240] loss: 0.447 [2, 300] loss: 0.436 [2, 360] loss: 0.434 Epoch: 2 -> Loss: 0.409345060587 Epoch: 2 -> Test Accuracy: 81.97 [3, 60] loss: 0.406 [3, 120] loss: 0.418 [3, 180] loss: 0.387 [3, 240] loss: 0.396 [3, 300] loss: 0.388 [3, 360] loss: 0.403 Epoch: 3 -> Loss: 0.35332608223 Epoch: 3 -> Test Accuracy: 83.44 [4, 60] loss: 0.362 [4, 120] loss: 0.360 [4, 180] loss: 0.363 [4, 240] loss: 0.370 [4, 300] loss: 0.377 [4, 360] loss: 0.376 Epoch: 4 -> Loss: 0.392022073269 Epoch: 4 -> Test Accuracy: 83.95 [5, 60] loss: 0.338 [5, 120] loss: 0.350 [5, 180] loss: 0.347 [5, 240] loss: 0.348 [5, 300] loss: 0.337 [5, 360] loss: 0.361 Epoch: 5 -> Loss: 0.339683711529 Epoch: 5 -> Test Accuracy: 83.27 [6, 60] loss: 0.325 [6, 120] loss: 0.344 [6, 180] loss: 0.339 [6, 240] loss: 0.342 [6, 300] loss: 0.350 [6, 360] loss: 0.322 Epoch: 6 -> Loss: 0.357249289751 Epoch: 6 -> Test Accuracy: 84.05 [7, 60] loss: 0.301 [7, 120] loss: 0.341 [7, 180] loss: 0.335 [7, 240] loss: 0.321 [7, 300] loss: 0.316 [7, 360] loss: 0.320 Epoch: 7 -> Loss: 0.389253050089 Epoch: 7 -> Test Accuracy: 84.71 [8, 60] loss: 0.293 [8, 120] loss: 0.294 [8, 180] loss: 0.304 [8, 240] loss: 0.317 [8, 300] loss: 0.332 [8, 360] loss: 0.335 Epoch: 8 -> Loss: 0.474091142416 Epoch: 8 -> Test Accuracy: 84.65 [9, 60] loss: 0.288 [9, 120] loss: 0.294 [9, 180] loss: 0.309 [9, 240] loss: 0.317 [9, 300] loss: 0.311 [9, 360] loss: 0.336 Epoch: 9 -> Loss: 0.41759377718 Epoch: 9 -> Test Accuracy: 85.32 [10, 60] loss: 0.285 [10, 120] loss: 0.290 [10, 180] loss: 0.296 [10, 240] loss: 0.316 [10, 300] loss: 0.310 [10, 360] loss: 0.316 Epoch: 10 -> Loss: 0.338132470846 Epoch: 10 -> Test Accuracy: 84.9 [11, 60] loss: 0.277 [11, 120] loss: 0.300 [11, 180] loss: 0.285 [11, 240] loss: 0.319 [11, 300] loss: 0.296 [11, 360] loss: 0.308 Epoch: 11 -> Loss: 0.167863562703 Epoch: 11 -> Test Accuracy: 85.73 [12, 60] loss: 0.262 [12, 120] loss: 0.277 [12, 180] loss: 0.289 [12, 240] loss: 0.327 [12, 300] loss: 0.299 [12, 360] loss: 0.292 Epoch: 12 -> Loss: 0.111307814717 Epoch: 12 -> Test Accuracy: 84.76 [13, 60] loss: 0.264 [13, 120] loss: 0.288 [13, 180] loss: 0.283 [13, 240] loss: 0.299 [13, 300] loss: 0.300 [13, 360] loss: 0.301 Epoch: 13 -> Loss: 0.252347409725 Epoch: 13 -> Test Accuracy: 85.5 [14, 60] loss: 0.260 [14, 120] loss: 0.279 [14, 180] loss: 0.284 [14, 240] loss: 0.301 [14, 300] loss: 0.297 [14, 360] loss: 0.294 Epoch: 14 -> Loss: 0.364129632711 Epoch: 14 -> Test Accuracy: 84.46 [15, 60] loss: 0.268 [15, 120] loss: 0.272 [15, 180] loss: 0.284 [15, 240] loss: 0.295 [15, 300] loss: 0.287 [15, 360] loss: 0.296 Epoch: 15 -> Loss: 0.400423526764 Epoch: 15 -> Test Accuracy: 84.29 [16, 60] loss: 0.255 [16, 120] loss: 0.280 [16, 180] loss: 0.272 [16, 240] loss: 0.281 [16, 300] loss: 0.276 [16, 360] loss: 0.303 Epoch: 16 -> Loss: 0.329053431749 Epoch: 16 -> Test Accuracy: 85.5 [17, 60] loss: 0.263 [17, 120] loss: 0.268 [17, 180] loss: 0.287 [17, 240] loss: 0.278 [17, 300] loss: 0.294 [17, 360] loss: 0.287 Epoch: 17 -> Loss: 0.170243829489 Epoch: 17 -> Test Accuracy: 85.22 [18, 60] loss: 0.253 [18, 120] loss: 0.269 [18, 180] loss: 0.283 [18, 240] loss: 0.272 [18, 300] loss: 0.280 [18, 360] loss: 0.291 Epoch: 18 -> Loss: 0.35972815752 Epoch: 18 -> Test Accuracy: 84.89 [19, 60] loss: 0.256 [19, 120] loss: 0.259 [19, 180] loss: 0.262 [19, 240] loss: 0.292 [19, 300] loss: 0.287 [19, 360] loss: 0.278 Epoch: 19 -> Loss: 0.249232262373 Epoch: 19 -> Test Accuracy: 85.28 [20, 60] loss: 0.262 [20, 120] loss: 0.264 [20, 180] loss: 0.276 [20, 240] loss: 0.286 [20, 300] loss: 0.293 [20, 360] loss: 0.278 Epoch: 20 -> Loss: 0.347967594862 Epoch: 20 -> Test Accuracy: 85.64 [21, 60] loss: 0.253 [21, 120] loss: 0.248 [21, 180] loss: 0.279 [21, 240] loss: 0.271 [21, 300] loss: 0.285 [21, 360] loss: 0.283 Epoch: 21 -> Loss: 0.398791998625 Epoch: 21 -> Test Accuracy: 86.31 [22, 60] loss: 0.245 [22, 120] loss: 0.239 [22, 180] loss: 0.264 [22, 240] loss: 0.284 [22, 300] loss: 0.263 [22, 360] loss: 0.289 Epoch: 22 -> Loss: 0.446104854345 Epoch: 22 -> Test Accuracy: 85.01 [23, 60] loss: 0.250 [23, 120] loss: 0.256 [23, 180] loss: 0.272 [23, 240] loss: 0.287 [23, 300] loss: 0.289 [23, 360] loss: 0.280 Epoch: 23 -> Loss: 0.238490581512 Epoch: 23 -> Test Accuracy: 84.32 [24, 60] loss: 0.240 [24, 120] loss: 0.261 [24, 180] loss: 0.262 [24, 240] loss: 0.270 [24, 300] loss: 0.286 [24, 360] loss: 0.275 Epoch: 24 -> Loss: 0.255210250616 Epoch: 24 -> Test Accuracy: 85.9 [25, 60] loss: 0.247 [25, 120] loss: 0.258 [25, 180] loss: 0.276 [25, 240] loss: 0.271 [25, 300] loss: 0.284 [25, 360] loss: 0.284 Epoch: 25 -> Loss: 0.337270259857 Epoch: 25 -> Test Accuracy: 85.78 [26, 60] loss: 0.247 [26, 120] loss: 0.240 [26, 180] loss: 0.259 [26, 240] loss: 0.263 [26, 300] loss: 0.269 [26, 360] loss: 0.301 Epoch: 26 -> Loss: 0.226263910532 Epoch: 26 -> Test Accuracy: 85.14 [27, 60] loss: 0.245 [27, 120] loss: 0.253 [27, 180] loss: 0.262 [27, 240] loss: 0.261 [27, 300] loss: 0.274 [27, 360] loss: 0.263 Epoch: 27 -> Loss: 0.300476163626 Epoch: 27 -> Test Accuracy: 85.61 [28, 60] loss: 0.253 [28, 120] loss: 0.252 [28, 180] loss: 0.268 [28, 240] loss: 0.268 [28, 300] loss: 0.254 [28, 360] loss: 0.278 Epoch: 28 -> Loss: 0.455658853054 Epoch: 28 -> Test Accuracy: 85.36 [29, 60] loss: 0.241 [29, 120] loss: 0.248 [29, 180] loss: 0.258 [29, 240] loss: 0.259 [29, 300] loss: 0.266 [29, 360] loss: 0.289 Epoch: 29 -> Loss: 0.295382708311 Epoch: 29 -> Test Accuracy: 84.71 [30, 60] loss: 0.233 [30, 120] loss: 0.243 [30, 180] loss: 0.262 [30, 240] loss: 0.266 [30, 300] loss: 0.294 [30, 360] loss: 0.291 Epoch: 30 -> Loss: 0.398016661406 Epoch: 30 -> Test Accuracy: 85.43 [31, 60] loss: 0.225 [31, 120] loss: 0.252 [31, 180] loss: 0.258 [31, 240] loss: 0.271 [31, 300] loss: 0.279 [31, 360] loss: 0.273 Epoch: 31 -> Loss: 0.36989736557 Epoch: 31 -> Test Accuracy: 85.47 [32, 60] loss: 0.252 [32, 120] loss: 0.250 [32, 180] loss: 0.263 [32, 240] loss: 0.261 [32, 300] loss: 0.268 [32, 360] loss: 0.285 Epoch: 32 -> Loss: 0.167249187827 Epoch: 32 -> Test Accuracy: 85.71 [33, 60] loss: 0.224 [33, 120] loss: 0.245 [33, 180] loss: 0.255 [33, 240] loss: 0.273 [33, 300] loss: 0.268 [33, 360] loss: 0.285 Epoch: 33 -> Loss: 0.213450461626 Epoch: 33 -> Test Accuracy: 85.77 [34, 60] loss: 0.237 [34, 120] loss: 0.238 [34, 180] loss: 0.263 [34, 240] loss: 0.264 [34, 300] loss: 0.266 [34, 360] loss: 0.274 Epoch: 34 -> Loss: 0.227099820971 Epoch: 34 -> Test Accuracy: 84.46 [35, 60] loss: 0.242 [35, 120] loss: 0.243 [35, 180] loss: 0.260 [35, 240] loss: 0.254 [35, 300] loss: 0.281 [35, 360] loss: 0.285 Epoch: 35 -> Loss: 0.20040102303 Epoch: 35 -> Test Accuracy: 85.7 [36, 60] loss: 0.207 [36, 120] loss: 0.180 [36, 180] loss: 0.163 [36, 240] loss: 0.174 [36, 300] loss: 0.161 [36, 360] loss: 0.168 Epoch: 36 -> Loss: 0.178841218352 Epoch: 36 -> Test Accuracy: 88.54 [37, 60] loss: 0.141 [37, 120] loss: 0.140 [37, 180] loss: 0.151 [37, 240] loss: 0.141 [37, 300] loss: 0.140 [37, 360] loss: 0.148 Epoch: 37 -> Loss: 0.19337554276 Epoch: 37 -> Test Accuracy: 87.98 [38, 60] loss: 0.130 [38, 120] loss: 0.121 [38, 180] loss: 0.136 [38, 240] loss: 0.130 [38, 300] loss: 0.133 [38, 360] loss: 0.144 Epoch: 38 -> Loss: 0.13021209836 Epoch: 38 -> Test Accuracy: 88.26 [39, 60] loss: 0.115 [39, 120] loss: 0.115 [39, 180] loss: 0.125 [39, 240] loss: 0.125 [39, 300] loss: 0.119 [39, 360] loss: 0.126 Epoch: 39 -> Loss: 0.205182701349 Epoch: 39 -> Test Accuracy: 88.46 [40, 60] loss: 0.109 [40, 120] loss: 0.114 [40, 180] loss: 0.120 [40, 240] loss: 0.116 [40, 300] loss: 0.123 [40, 360] loss: 0.116 Epoch: 40 -> Loss: 0.126770943403 Epoch: 40 -> Test Accuracy: 87.99 [41, 60] loss: 0.112 [41, 120] loss: 0.110 [41, 180] loss: 0.112 [41, 240] loss: 0.115 [41, 300] loss: 0.107 [41, 360] loss: 0.114 Epoch: 41 -> Loss: 0.173801928759 Epoch: 41 -> Test Accuracy: 88.23 [42, 60] loss: 0.102 [42, 120] loss: 0.107 [42, 180] loss: 0.109 [42, 240] loss: 0.109 [42, 300] loss: 0.110 [42, 360] loss: 0.111 Epoch: 42 -> Loss: 0.237250089645 Epoch: 42 -> Test Accuracy: 87.73 [43, 60] loss: 0.098 [43, 120] loss: 0.100 [43, 180] loss: 0.107 [43, 240] loss: 0.104 [43, 300] loss: 0.111 [43, 360] loss: 0.115 Epoch: 43 -> Loss: 0.16972528398 Epoch: 43 -> Test Accuracy: 87.6 [44, 60] loss: 0.094 [44, 120] loss: 0.104 [44, 180] loss: 0.106 [44, 240] loss: 0.110 [44, 300] loss: 0.112 [44, 360] loss: 0.119 Epoch: 44 -> Loss: 0.175173729658 Epoch: 44 -> Test Accuracy: 87.76 [45, 60] loss: 0.095 [45, 120] loss: 0.100 [45, 180] loss: 0.098 [45, 240] loss: 0.102 [45, 300] loss: 0.105 [45, 360] loss: 0.112 Epoch: 45 -> Loss: 0.245339676738 Epoch: 45 -> Test Accuracy: 87.54 [46, 60] loss: 0.098 [46, 120] loss: 0.096 [46, 180] loss: 0.101 [46, 240] loss: 0.100 [46, 300] loss: 0.113 [46, 360] loss: 0.108 Epoch: 46 -> Loss: 0.222394153476 Epoch: 46 -> Test Accuracy: 87.71 [47, 60] loss: 0.098 [47, 120] loss: 0.105 [47, 180] loss: 0.104 [47, 240] loss: 0.103 [47, 300] loss: 0.102 [47, 360] loss: 0.112 Epoch: 47 -> Loss: 0.197726413608 Epoch: 47 -> Test Accuracy: 87.08 [48, 60] loss: 0.099 [48, 120] loss: 0.098 [48, 180] loss: 0.103 [48, 240] loss: 0.110 [48, 300] loss: 0.112 [48, 360] loss: 0.113 Epoch: 48 -> Loss: 0.312668889761 Epoch: 48 -> Test Accuracy: 87.26 [49, 60] loss: 0.087 [49, 120] loss: 0.103 [49, 180] loss: 0.108 [49, 240] loss: 0.108 [49, 300] loss: 0.113 [49, 360] loss: 0.109 Epoch: 49 -> Loss: 0.146272867918 Epoch: 49 -> Test Accuracy: 87.92 [50, 60] loss: 0.095 [50, 120] loss: 0.097 [50, 180] loss: 0.111 [50, 240] loss: 0.108 [50, 300] loss: 0.118 [50, 360] loss: 0.111 Epoch: 50 -> Loss: 0.1028393507 Epoch: 50 -> Test Accuracy: 87.6 [51, 60] loss: 0.092 [51, 120] loss: 0.096 [51, 180] loss: 0.104 [51, 240] loss: 0.102 [51, 300] loss: 0.120 [51, 360] loss: 0.113 Epoch: 51 -> Loss: 0.191755786538 Epoch: 51 -> Test Accuracy: 87.34 [52, 60] loss: 0.103 [52, 120] loss: 0.105 [52, 180] loss: 0.105 [52, 240] loss: 0.108 [52, 300] loss: 0.103 [52, 360] loss: 0.103 Epoch: 52 -> Loss: 0.156757131219 Epoch: 52 -> Test Accuracy: 87.54 [53, 60] loss: 0.096 [53, 120] loss: 0.107 [53, 180] loss: 0.109 [53, 240] loss: 0.110 [53, 300] loss: 0.106 [53, 360] loss: 0.116 Epoch: 53 -> Loss: 0.137719243765 Epoch: 53 -> Test Accuracy: 87.48 [54, 60] loss: 0.093 [54, 120] loss: 0.105 [54, 180] loss: 0.108 [54, 240] loss: 0.115 [54, 300] loss: 0.107 [54, 360] loss: 0.115 Epoch: 54 -> Loss: 0.1882378757 Epoch: 54 -> Test Accuracy: 87.12 [55, 60] loss: 0.091 [55, 120] loss: 0.100 [55, 180] loss: 0.100 [55, 240] loss: 0.102 [55, 300] loss: 0.109 [55, 360] loss: 0.124 Epoch: 55 -> Loss: 0.154113784432 Epoch: 55 -> Test Accuracy: 86.94 [56, 60] loss: 0.097 [56, 120] loss: 0.104 [56, 180] loss: 0.106 [56, 240] loss: 0.112 [56, 300] loss: 0.117 [56, 360] loss: 0.105 Epoch: 56 -> Loss: 0.191437244415 Epoch: 56 -> Test Accuracy: 86.89 [57, 60] loss: 0.092 [57, 120] loss: 0.096 [57, 180] loss: 0.100 [57, 240] loss: 0.108 [57, 300] loss: 0.118 [57, 360] loss: 0.119 Epoch: 57 -> Loss: 0.182171300054 Epoch: 57 -> Test Accuracy: 86.83 [58, 60] loss: 0.104 [58, 120] loss: 0.099 [58, 180] loss: 0.098 [58, 240] loss: 0.105 [58, 300] loss: 0.103 [58, 360] loss: 0.126 Epoch: 58 -> Loss: 0.0771923214197 Epoch: 58 -> Test Accuracy: 87.42 [59, 60] loss: 0.099 [59, 120] loss: 0.098 [59, 180] loss: 0.106 [59, 240] loss: 0.112 [59, 300] loss: 0.113 [59, 360] loss: 0.107 Epoch: 59 -> Loss: 0.108989581466 Epoch: 59 -> Test Accuracy: 87.63 [60, 60] loss: 0.101 [60, 120] loss: 0.094 [60, 180] loss: 0.095 [60, 240] loss: 0.102 [60, 300] loss: 0.110 [60, 360] loss: 0.118 Epoch: 60 -> Loss: 0.125832885504 Epoch: 60 -> Test Accuracy: 87.84 [61, 60] loss: 0.096 [61, 120] loss: 0.100 [61, 180] loss: 0.105 [61, 240] loss: 0.093 [61, 300] loss: 0.112 [61, 360] loss: 0.114 Epoch: 61 -> Loss: 0.0430082045496 Epoch: 61 -> Test Accuracy: 86.99 [62, 60] loss: 0.094 [62, 120] loss: 0.096 [62, 180] loss: 0.110 [62, 240] loss: 0.106 [62, 300] loss: 0.112 [62, 360] loss: 0.111 Epoch: 62 -> Loss: 0.100391790271 Epoch: 62 -> Test Accuracy: 87.25 [63, 60] loss: 0.098 [63, 120] loss: 0.100 [63, 180] loss: 0.099 [63, 240] loss: 0.102 [63, 300] loss: 0.102 [63, 360] loss: 0.113 Epoch: 63 -> Loss: 0.0969136804342 Epoch: 63 -> Test Accuracy: 87.5 [64, 60] loss: 0.093 [64, 120] loss: 0.104 [64, 180] loss: 0.095 [64, 240] loss: 0.105 [64, 300] loss: 0.101 [64, 360] loss: 0.106 Epoch: 64 -> Loss: 0.106285773218 Epoch: 64 -> Test Accuracy: 86.7 [65, 60] loss: 0.090 [65, 120] loss: 0.098 [65, 180] loss: 0.105 [65, 240] loss: 0.093 [65, 300] loss: 0.099 [65, 360] loss: 0.109 Epoch: 65 -> Loss: 0.0558515302837 Epoch: 65 -> Test Accuracy: 87.15 [66, 60] loss: 0.100 [66, 120] loss: 0.091 [66, 180] loss: 0.105 [66, 240] loss: 0.101 [66, 300] loss: 0.103 [66, 360] loss: 0.118 Epoch: 66 -> Loss: 0.10032980144 Epoch: 66 -> Test Accuracy: 87.0 [67, 60] loss: 0.097 [67, 120] loss: 0.098 [67, 180] loss: 0.089 [67, 240] loss: 0.101 [67, 300] loss: 0.105 [67, 360] loss: 0.113 Epoch: 67 -> Loss: 0.105195082724 Epoch: 67 -> Test Accuracy: 87.18 [68, 60] loss: 0.097 [68, 120] loss: 0.098 [68, 180] loss: 0.096 [68, 240] loss: 0.104 [68, 300] loss: 0.094 [68, 360] loss: 0.106 Epoch: 68 -> Loss: 0.166546106339 Epoch: 68 -> Test Accuracy: 87.57 [69, 60] loss: 0.093 [69, 120] loss: 0.095 [69, 180] loss: 0.089 [69, 240] loss: 0.114 [69, 300] loss: 0.113 [69, 360] loss: 0.114 Epoch: 69 -> Loss: 0.0638231262565 Epoch: 69 -> Test Accuracy: 86.86 [70, 60] loss: 0.099 [70, 120] loss: 0.098 [70, 180] loss: 0.098 [70, 240] loss: 0.107 [70, 300] loss: 0.112 [70, 360] loss: 0.110 Epoch: 70 -> Loss: 0.181370839477 Epoch: 70 -> Test Accuracy: 87.54 [71, 60] loss: 0.079 [71, 120] loss: 0.070 [71, 180] loss: 0.068 [71, 240] loss: 0.060 [71, 300] loss: 0.059 [71, 360] loss: 0.057 Epoch: 71 -> Loss: 0.129561260343 Epoch: 71 -> Test Accuracy: 88.99 [72, 60] loss: 0.053 [72, 120] loss: 0.052 [72, 180] loss: 0.053 [72, 240] loss: 0.047 [72, 300] loss: 0.051 [72, 360] loss: 0.050 Epoch: 72 -> Loss: 0.0847681388259 Epoch: 72 -> Test Accuracy: 88.93 [73, 60] loss: 0.047 [73, 120] loss: 0.044 [73, 180] loss: 0.049 [73, 240] loss: 0.047 [73, 300] loss: 0.048 [73, 360] loss: 0.049 Epoch: 73 -> Loss: 0.113650344312 Epoch: 73 -> Test Accuracy: 89.19 [74, 60] loss: 0.045 [74, 120] loss: 0.046 [74, 180] loss: 0.041 [74, 240] loss: 0.044 [74, 300] loss: 0.045 [74, 360] loss: 0.041 Epoch: 74 -> Loss: 0.0344423241913 Epoch: 74 -> Test Accuracy: 88.99 [75, 60] loss: 0.041 [75, 120] loss: 0.045 [75, 180] loss: 0.039 [75, 240] loss: 0.041 [75, 300] loss: 0.042 [75, 360] loss: 0.041 Epoch: 75 -> Loss: 0.0303840301931 Epoch: 75 -> Test Accuracy: 88.75 [76, 60] loss: 0.038 [76, 120] loss: 0.040 [76, 180] loss: 0.041 [76, 240] loss: 0.041 [76, 300] loss: 0.041 [76, 360] loss: 0.036 Epoch: 76 -> Loss: 0.0474015362561 Epoch: 76 -> Test Accuracy: 88.83 [77, 60] loss: 0.037 [77, 120] loss: 0.037 [77, 180] loss: 0.036 [77, 240] loss: 0.036 [77, 300] loss: 0.039 [77, 360] loss: 0.040 Epoch: 77 -> Loss: 0.0596053190529 Epoch: 77 -> Test Accuracy: 88.81 [78, 60] loss: 0.035 [78, 120] loss: 0.035 [78, 180] loss: 0.038 [78, 240] loss: 0.036 [78, 300] loss: 0.041 [78, 360] loss: 0.038 Epoch: 78 -> Loss: 0.0439391285181 Epoch: 78 -> Test Accuracy: 88.74 [79, 60] loss: 0.034 [79, 120] loss: 0.034 [79, 180] loss: 0.035 [79, 240] loss: 0.035 [79, 300] loss: 0.037 [79, 360] loss: 0.036 Epoch: 79 -> Loss: 0.044279973954 Epoch: 79 -> Test Accuracy: 88.83 [80, 60] loss: 0.034 [80, 120] loss: 0.034 [80, 180] loss: 0.032 [80, 240] loss: 0.034 [80, 300] loss: 0.035 [80, 360] loss: 0.038 Epoch: 80 -> Loss: 0.0172802563757 Epoch: 80 -> Test Accuracy: 88.64 [81, 60] loss: 0.033 [81, 120] loss: 0.031 [81, 180] loss: 0.031 [81, 240] loss: 0.033 [81, 300] loss: 0.031 [81, 360] loss: 0.037 Epoch: 81 -> Loss: 0.026349067688 Epoch: 81 -> Test Accuracy: 88.83 [82, 60] loss: 0.031 [82, 120] loss: 0.031 [82, 180] loss: 0.032 [82, 240] loss: 0.034 [82, 300] loss: 0.033 [82, 360] loss: 0.030 Epoch: 82 -> Loss: 0.020774345845 Epoch: 82 -> Test Accuracy: 88.91 [83, 60] loss: 0.029 [83, 120] loss: 0.030 [83, 180] loss: 0.031 [83, 240] loss: 0.033 [83, 300] loss: 0.033 [83, 360] loss: 0.037 Epoch: 83 -> Loss: 0.015046775341 Epoch: 83 -> Test Accuracy: 88.94 [84, 60] loss: 0.029 [84, 120] loss: 0.030 [84, 180] loss: 0.030 [84, 240] loss: 0.033 [84, 300] loss: 0.031 [84, 360] loss: 0.031 Epoch: 84 -> Loss: 0.0233301632106 Epoch: 84 -> Test Accuracy: 88.85 [85, 60] loss: 0.029 [85, 120] loss: 0.030 [85, 180] loss: 0.030 [85, 240] loss: 0.031 [85, 300] loss: 0.031 [85, 360] loss: 0.031 Epoch: 85 -> Loss: 0.0523966550827 Epoch: 85 -> Test Accuracy: 88.72 [86, 60] loss: 0.028 [86, 120] loss: 0.029 [86, 180] loss: 0.030 [86, 240] loss: 0.027 [86, 300] loss: 0.030 [86, 360] loss: 0.029 Epoch: 86 -> Loss: 0.0145873129368 Epoch: 86 -> Test Accuracy: 88.98 [87, 60] loss: 0.027 [87, 120] loss: 0.026 [87, 180] loss: 0.024 [87, 240] loss: 0.028 [87, 300] loss: 0.027 [87, 360] loss: 0.025 Epoch: 87 -> Loss: 0.025675997138 Epoch: 87 -> Test Accuracy: 88.84 [88, 60] loss: 0.028 [88, 120] loss: 0.028 [88, 180] loss: 0.028 [88, 240] loss: 0.025 [88, 300] loss: 0.027 [88, 360] loss: 0.026 Epoch: 88 -> Loss: 0.0189745239913 Epoch: 88 -> Test Accuracy: 88.86 [89, 60] loss: 0.026 [89, 120] loss: 0.027 [89, 180] loss: 0.025 [89, 240] loss: 0.027 [89, 300] loss: 0.025 [89, 360] loss: 0.029 Epoch: 89 -> Loss: 0.0349076017737 Epoch: 89 -> Test Accuracy: 88.94 [90, 60] loss: 0.025 [90, 120] loss: 0.026 [90, 180] loss: 0.026 [90, 240] loss: 0.024 [90, 300] loss: 0.028 [90, 360] loss: 0.025 Epoch: 90 -> Loss: 0.046495012939 Epoch: 90 -> Test Accuracy: 88.82 [91, 60] loss: 0.025 [91, 120] loss: 0.026 [91, 180] loss: 0.026 [91, 240] loss: 0.027 [91, 300] loss: 0.024 [91, 360] loss: 0.024 Epoch: 91 -> Loss: 0.0370255708694 Epoch: 91 -> Test Accuracy: 88.99 [92, 60] loss: 0.024 [92, 120] loss: 0.024 [92, 180] loss: 0.024 [92, 240] loss: 0.025 [92, 300] loss: 0.027 [92, 360] loss: 0.026 Epoch: 92 -> Loss: 0.0183320045471 Epoch: 92 -> Test Accuracy: 88.91 [93, 60] loss: 0.025 [93, 120] loss: 0.025 [93, 180] loss: 0.023 [93, 240] loss: 0.024 [93, 300] loss: 0.025 [93, 360] loss: 0.026 Epoch: 93 -> Loss: 0.0449894368649 Epoch: 93 -> Test Accuracy: 88.89 [94, 60] loss: 0.024 [94, 120] loss: 0.026 [94, 180] loss: 0.025 [94, 240] loss: 0.026 [94, 300] loss: 0.026 [94, 360] loss: 0.026 Epoch: 94 -> Loss: 0.0578770749271 Epoch: 94 -> Test Accuracy: 88.83 [95, 60] loss: 0.024 [95, 120] loss: 0.026 [95, 180] loss: 0.026 [95, 240] loss: 0.023 [95, 300] loss: 0.025 [95, 360] loss: 0.026 Epoch: 95 -> Loss: 0.0535115115345 Epoch: 95 -> Test Accuracy: 88.78 [96, 60] loss: 0.025 [96, 120] loss: 0.026 [96, 180] loss: 0.026 [96, 240] loss: 0.025 [96, 300] loss: 0.023 [96, 360] loss: 0.026 Epoch: 96 -> Loss: 0.0392332822084 Epoch: 96 -> Test Accuracy: 88.87 [97, 60] loss: 0.026 [97, 120] loss: 0.025 [97, 180] loss: 0.027 [97, 240] loss: 0.025 [97, 300] loss: 0.027 [97, 360] loss: 0.027 Epoch: 97 -> Loss: 0.0249798540026 Epoch: 97 -> Test Accuracy: 88.83 [98, 60] loss: 0.026 [98, 120] loss: 0.025 [98, 180] loss: 0.026 [98, 240] loss: 0.027 [98, 300] loss: 0.024 [98, 360] loss: 0.024 Epoch: 98 -> Loss: 0.0612790510058 Epoch: 98 -> Test Accuracy: 89.01 [99, 60] loss: 0.025 [99, 120] loss: 0.024 [99, 180] loss: 0.024 [99, 240] loss: 0.024 [99, 300] loss: 0.025 [99, 360] loss: 0.025 Epoch: 99 -> Loss: 0.023496394977 Epoch: 99 -> Test Accuracy: 89.03 [100, 60] loss: 0.025 [100, 120] loss: 0.024 [100, 180] loss: 0.025 [100, 240] loss: 0.022 [100, 300] loss: 0.024 [100, 360] loss: 0.025 Epoch: 100 -> Loss: 0.0243260748684 Epoch: 100 -> Test Accuracy: 89.0 Finished Training [1, 60] loss: 0.889 [1, 120] loss: 0.654 [1, 180] loss: 0.592 [1, 240] loss: 0.572 [1, 300] loss: 0.549 [1, 360] loss: 0.529 Epoch: 1 -> Loss: 0.508641719818 Epoch: 1 -> Test Accuracy: 78.19 [2, 60] loss: 0.504 [2, 120] loss: 0.491 [2, 180] loss: 0.490 [2, 240] loss: 0.500 [2, 300] loss: 0.486 [2, 360] loss: 0.462 Epoch: 2 -> Loss: 0.384296238422 Epoch: 2 -> Test Accuracy: 80.43 [3, 60] loss: 0.456 [3, 120] loss: 0.444 [3, 180] loss: 0.467 [3, 240] loss: 0.450 [3, 300] loss: 0.457 [3, 360] loss: 0.465 Epoch: 3 -> Loss: 0.424328237772 Epoch: 3 -> Test Accuracy: 80.89 [4, 60] loss: 0.425 [4, 120] loss: 0.445 [4, 180] loss: 0.446 [4, 240] loss: 0.425 [4, 300] loss: 0.439 [4, 360] loss: 0.433 Epoch: 4 -> Loss: 0.433434545994 Epoch: 4 -> Test Accuracy: 80.85 [5, 60] loss: 0.407 [5, 120] loss: 0.410 [5, 180] loss: 0.428 [5, 240] loss: 0.416 [5, 300] loss: 0.431 [5, 360] loss: 0.409 Epoch: 5 -> Loss: 0.510099947453 Epoch: 5 -> Test Accuracy: 80.3 [6, 60] loss: 0.388 [6, 120] loss: 0.412 [6, 180] loss: 0.408 [6, 240] loss: 0.419 [6, 300] loss: 0.409 [6, 360] loss: 0.427 Epoch: 6 -> Loss: 0.52281999588 Epoch: 6 -> Test Accuracy: 81.66 [7, 60] loss: 0.385 [7, 120] loss: 0.390 [7, 180] loss: 0.401 [7, 240] loss: 0.405 [7, 300] loss: 0.405 [7, 360] loss: 0.413 Epoch: 7 -> Loss: 0.556824326515 Epoch: 7 -> Test Accuracy: 82.14 [8, 60] loss: 0.380 [8, 120] loss: 0.389 [8, 180] loss: 0.399 [8, 240] loss: 0.403 [8, 300] loss: 0.385 [8, 360] loss: 0.391 Epoch: 8 -> Loss: 0.341546297073 Epoch: 8 -> Test Accuracy: 82.13 [9, 60] loss: 0.375 [9, 120] loss: 0.389 [9, 180] loss: 0.380 [9, 240] loss: 0.383 [9, 300] loss: 0.404 [9, 360] loss: 0.411 Epoch: 9 -> Loss: 0.266939967871 Epoch: 9 -> Test Accuracy: 82.25 [10, 60] loss: 0.366 [10, 120] loss: 0.365 [10, 180] loss: 0.406 [10, 240] loss: 0.377 [10, 300] loss: 0.389 [10, 360] loss: 0.404 Epoch: 10 -> Loss: 0.356192320585 Epoch: 10 -> Test Accuracy: 82.64 [11, 60] loss: 0.365 [11, 120] loss: 0.394 [11, 180] loss: 0.374 [11, 240] loss: 0.365 [11, 300] loss: 0.395 [11, 360] loss: 0.370 Epoch: 11 -> Loss: 0.549596965313 Epoch: 11 -> Test Accuracy: 82.38 [12, 60] loss: 0.351 [12, 120] loss: 0.371 [12, 180] loss: 0.364 [12, 240] loss: 0.376 [12, 300] loss: 0.367 [12, 360] loss: 0.387 Epoch: 12 -> Loss: 0.352015554905 Epoch: 12 -> Test Accuracy: 81.7 [13, 60] loss: 0.376 [13, 120] loss: 0.362 [13, 180] loss: 0.367 [13, 240] loss: 0.375 [13, 300] loss: 0.399 [13, 360] loss: 0.379 Epoch: 13 -> Loss: 0.467985540628 Epoch: 13 -> Test Accuracy: 82.17 [14, 60] loss: 0.356 [14, 120] loss: 0.358 [14, 180] loss: 0.360 [14, 240] loss: 0.366 [14, 300] loss: 0.383 [14, 360] loss: 0.393 Epoch: 14 -> Loss: 0.297144144773 Epoch: 14 -> Test Accuracy: 82.74 [15, 60] loss: 0.344 [15, 120] loss: 0.356 [15, 180] loss: 0.362 [15, 240] loss: 0.383 [15, 300] loss: 0.366 [15, 360] loss: 0.395 Epoch: 15 -> Loss: 0.327633023262 Epoch: 15 -> Test Accuracy: 81.06 [16, 60] loss: 0.361 [16, 120] loss: 0.348 [16, 180] loss: 0.371 [16, 240] loss: 0.363 [16, 300] loss: 0.369 [16, 360] loss: 0.376 Epoch: 16 -> Loss: 0.45045799017 Epoch: 16 -> Test Accuracy: 82.44 [17, 60] loss: 0.339 [17, 120] loss: 0.366 [17, 180] loss: 0.362 [17, 240] loss: 0.360 [17, 300] loss: 0.372 [17, 360] loss: 0.378 Epoch: 17 -> Loss: 0.355867266655 Epoch: 17 -> Test Accuracy: 82.02 [18, 60] loss: 0.351 [18, 120] loss: 0.353 [18, 180] loss: 0.358 [18, 240] loss: 0.362 [18, 300] loss: 0.384 [18, 360] loss: 0.377 Epoch: 18 -> Loss: 0.317616194487 Epoch: 18 -> Test Accuracy: 81.78 [19, 60] loss: 0.338 [19, 120] loss: 0.361 [19, 180] loss: 0.359 [19, 240] loss: 0.351 [19, 300] loss: 0.366 [19, 360] loss: 0.381 Epoch: 19 -> Loss: 0.389977753162 Epoch: 19 -> Test Accuracy: 82.57 [20, 60] loss: 0.357 [20, 120] loss: 0.340 [20, 180] loss: 0.367 [20, 240] loss: 0.355 [20, 300] loss: 0.369 [20, 360] loss: 0.348 Epoch: 20 -> Loss: 0.376787424088 Epoch: 20 -> Test Accuracy: 82.24 [21, 60] loss: 0.344 [21, 120] loss: 0.364 [21, 180] loss: 0.348 [21, 240] loss: 0.368 [21, 300] loss: 0.371 [21, 360] loss: 0.370 Epoch: 21 -> Loss: 0.313180625439 Epoch: 21 -> Test Accuracy: 82.93 [22, 60] loss: 0.340 [22, 120] loss: 0.361 [22, 180] loss: 0.348 [22, 240] loss: 0.363 [22, 300] loss: 0.360 [22, 360] loss: 0.348 Epoch: 22 -> Loss: 0.509989202023 Epoch: 22 -> Test Accuracy: 81.91 [23, 60] loss: 0.350 [23, 120] loss: 0.350 [23, 180] loss: 0.359 [23, 240] loss: 0.354 [23, 300] loss: 0.384 [23, 360] loss: 0.354 Epoch: 23 -> Loss: 0.430521190166 Epoch: 23 -> Test Accuracy: 82.29 [24, 60] loss: 0.330 [24, 120] loss: 0.347 [24, 180] loss: 0.364 [24, 240] loss: 0.359 [24, 300] loss: 0.361 [24, 360] loss: 0.365 Epoch: 24 -> Loss: 0.409759223461 Epoch: 24 -> Test Accuracy: 82.45 [25, 60] loss: 0.351 [25, 120] loss: 0.348 [25, 180] loss: 0.348 [25, 240] loss: 0.360 [25, 300] loss: 0.355 [25, 360] loss: 0.363 Epoch: 25 -> Loss: 0.385100871325 Epoch: 25 -> Test Accuracy: 81.57 [26, 60] loss: 0.343 [26, 120] loss: 0.347 [26, 180] loss: 0.358 [26, 240] loss: 0.359 [26, 300] loss: 0.359 [26, 360] loss: 0.355 Epoch: 26 -> Loss: 0.188759878278 Epoch: 26 -> Test Accuracy: 81.85 [27, 60] loss: 0.319 [27, 120] loss: 0.355 [27, 180] loss: 0.363 [27, 240] loss: 0.342 [27, 300] loss: 0.365 [27, 360] loss: 0.365 Epoch: 27 -> Loss: 0.308426380157 Epoch: 27 -> Test Accuracy: 83.0 [28, 60] loss: 0.322 [28, 120] loss: 0.353 [28, 180] loss: 0.352 [28, 240] loss: 0.361 [28, 300] loss: 0.360 [28, 360] loss: 0.367 Epoch: 28 -> Loss: 0.321726232767 Epoch: 28 -> Test Accuracy: 82.94 [29, 60] loss: 0.328 [29, 120] loss: 0.343 [29, 180] loss: 0.330 [29, 240] loss: 0.351 [29, 300] loss: 0.364 [29, 360] loss: 0.384 Epoch: 29 -> Loss: 0.480570882559 Epoch: 29 -> Test Accuracy: 83.13 [30, 60] loss: 0.344 [30, 120] loss: 0.338 [30, 180] loss: 0.347 [30, 240] loss: 0.345 [30, 300] loss: 0.355 [30, 360] loss: 0.375 Epoch: 30 -> Loss: 0.275167644024 Epoch: 30 -> Test Accuracy: 82.77 [31, 60] loss: 0.323 [31, 120] loss: 0.353 [31, 180] loss: 0.337 [31, 240] loss: 0.359 [31, 300] loss: 0.358 [31, 360] loss: 0.350 Epoch: 31 -> Loss: 0.261042743921 Epoch: 31 -> Test Accuracy: 82.52 [32, 60] loss: 0.337 [32, 120] loss: 0.344 [32, 180] loss: 0.331 [32, 240] loss: 0.350 [32, 300] loss: 0.354 [32, 360] loss: 0.355 Epoch: 32 -> Loss: 0.396286129951 Epoch: 32 -> Test Accuracy: 82.54 [33, 60] loss: 0.337 [33, 120] loss: 0.335 [33, 180] loss: 0.335 [33, 240] loss: 0.366 [33, 300] loss: 0.353 [33, 360] loss: 0.350 Epoch: 33 -> Loss: 0.399610996246 Epoch: 33 -> Test Accuracy: 83.0 [34, 60] loss: 0.320 [34, 120] loss: 0.343 [34, 180] loss: 0.348 [34, 240] loss: 0.349 [34, 300] loss: 0.365 [34, 360] loss: 0.353 Epoch: 34 -> Loss: 0.399015009403 Epoch: 34 -> Test Accuracy: 82.46 [35, 60] loss: 0.344 [35, 120] loss: 0.347 [35, 180] loss: 0.353 [35, 240] loss: 0.343 [35, 300] loss: 0.374 [35, 360] loss: 0.349 Epoch: 35 -> Loss: 0.489450037479 Epoch: 35 -> Test Accuracy: 82.84 [36, 60] loss: 0.287 [36, 120] loss: 0.259 [36, 180] loss: 0.265 [36, 240] loss: 0.267 [36, 300] loss: 0.257 [36, 360] loss: 0.268 Epoch: 36 -> Loss: 0.261972039938 Epoch: 36 -> Test Accuracy: 84.77 [37, 60] loss: 0.242 [37, 120] loss: 0.240 [37, 180] loss: 0.253 [37, 240] loss: 0.241 [37, 300] loss: 0.230 [37, 360] loss: 0.228 Epoch: 37 -> Loss: 0.270760238171 Epoch: 37 -> Test Accuracy: 84.92 [38, 60] loss: 0.217 [38, 120] loss: 0.226 [38, 180] loss: 0.216 [38, 240] loss: 0.242 [38, 300] loss: 0.236 [38, 360] loss: 0.233 Epoch: 38 -> Loss: 0.24845738709 Epoch: 38 -> Test Accuracy: 84.63 [39, 60] loss: 0.211 [39, 120] loss: 0.212 [39, 180] loss: 0.224 [39, 240] loss: 0.224 [39, 300] loss: 0.220 [39, 360] loss: 0.235 Epoch: 39 -> Loss: 0.275641053915 Epoch: 39 -> Test Accuracy: 84.84 [40, 60] loss: 0.219 [40, 120] loss: 0.204 [40, 180] loss: 0.227 [40, 240] loss: 0.219 [40, 300] loss: 0.216 [40, 360] loss: 0.226 Epoch: 40 -> Loss: 0.239670708776 Epoch: 40 -> Test Accuracy: 84.82 [41, 60] loss: 0.212 [41, 120] loss: 0.212 [41, 180] loss: 0.213 [41, 240] loss: 0.212 [41, 300] loss: 0.202 [41, 360] loss: 0.215 Epoch: 41 -> Loss: 0.135684221983 Epoch: 41 -> Test Accuracy: 84.71 [42, 60] loss: 0.200 [42, 120] loss: 0.207 [42, 180] loss: 0.208 [42, 240] loss: 0.208 [42, 300] loss: 0.221 [42, 360] loss: 0.213 Epoch: 42 -> Loss: 0.231472179294 Epoch: 42 -> Test Accuracy: 84.22 [43, 60] loss: 0.210 [43, 120] loss: 0.208 [43, 180] loss: 0.207 [43, 240] loss: 0.198 [43, 300] loss: 0.213 [43, 360] loss: 0.234 Epoch: 43 -> Loss: 0.147014141083 Epoch: 43 -> Test Accuracy: 84.49 [44, 60] loss: 0.202 [44, 120] loss: 0.205 [44, 180] loss: 0.194 [44, 240] loss: 0.210 [44, 300] loss: 0.212 [44, 360] loss: 0.218 Epoch: 44 -> Loss: 0.213748186827 Epoch: 44 -> Test Accuracy: 84.87 [45, 60] loss: 0.202 [45, 120] loss: 0.203 [45, 180] loss: 0.209 [45, 240] loss: 0.201 [45, 300] loss: 0.216 [45, 360] loss: 0.210 Epoch: 45 -> Loss: 0.192213818431 Epoch: 45 -> Test Accuracy: 84.53 [46, 60] loss: 0.183 [46, 120] loss: 0.193 [46, 180] loss: 0.219 [46, 240] loss: 0.206 [46, 300] loss: 0.211 [46, 360] loss: 0.211 Epoch: 46 -> Loss: 0.204710766673 Epoch: 46 -> Test Accuracy: 84.63 [47, 60] loss: 0.189 [47, 120] loss: 0.203 [47, 180] loss: 0.200 [47, 240] loss: 0.202 [47, 300] loss: 0.211 [47, 360] loss: 0.210 Epoch: 47 -> Loss: 0.170248359442 Epoch: 47 -> Test Accuracy: 84.02 [48, 60] loss: 0.188 [48, 120] loss: 0.198 [48, 180] loss: 0.197 [48, 240] loss: 0.201 [48, 300] loss: 0.207 [48, 360] loss: 0.227 Epoch: 48 -> Loss: 0.326770961285 Epoch: 48 -> Test Accuracy: 84.21 [49, 60] loss: 0.192 [49, 120] loss: 0.196 [49, 180] loss: 0.202 [49, 240] loss: 0.201 [49, 300] loss: 0.207 [49, 360] loss: 0.207 Epoch: 49 -> Loss: 0.268282830715 Epoch: 49 -> Test Accuracy: 84.24 [50, 60] loss: 0.189 [50, 120] loss: 0.203 [50, 180] loss: 0.216 [50, 240] loss: 0.199 [50, 300] loss: 0.204 [50, 360] loss: 0.210 Epoch: 50 -> Loss: 0.210638567805 Epoch: 50 -> Test Accuracy: 84.04 [51, 60] loss: 0.194 [51, 120] loss: 0.204 [51, 180] loss: 0.205 [51, 240] loss: 0.201 [51, 300] loss: 0.219 [51, 360] loss: 0.197 Epoch: 51 -> Loss: 0.383303076029 Epoch: 51 -> Test Accuracy: 84.26 [52, 60] loss: 0.193 [52, 120] loss: 0.192 [52, 180] loss: 0.200 [52, 240] loss: 0.209 [52, 300] loss: 0.191 [52, 360] loss: 0.208 Epoch: 52 -> Loss: 0.351854145527 Epoch: 52 -> Test Accuracy: 84.34 [53, 60] loss: 0.197 [53, 120] loss: 0.190 [53, 180] loss: 0.191 [53, 240] loss: 0.211 [53, 300] loss: 0.202 [53, 360] loss: 0.213 Epoch: 53 -> Loss: 0.374006152153 Epoch: 53 -> Test Accuracy: 84.17 [54, 60] loss: 0.195 [54, 120] loss: 0.195 [54, 180] loss: 0.206 [54, 240] loss: 0.207 [54, 300] loss: 0.186 [54, 360] loss: 0.209 Epoch: 54 -> Loss: 0.195482745767 Epoch: 54 -> Test Accuracy: 83.98 [55, 60] loss: 0.191 [55, 120] loss: 0.193 [55, 180] loss: 0.195 [55, 240] loss: 0.201 [55, 300] loss: 0.208 [55, 360] loss: 0.207 Epoch: 55 -> Loss: 0.172643795609 Epoch: 55 -> Test Accuracy: 83.59 [56, 60] loss: 0.185 [56, 120] loss: 0.187 [56, 180] loss: 0.196 [56, 240] loss: 0.197 [56, 300] loss: 0.205 [56, 360] loss: 0.210 Epoch: 56 -> Loss: 0.169902250171 Epoch: 56 -> Test Accuracy: 84.19 [57, 60] loss: 0.192 [57, 120] loss: 0.197 [57, 180] loss: 0.190 [57, 240] loss: 0.202 [57, 300] loss: 0.198 [57, 360] loss: 0.204 Epoch: 57 -> Loss: 0.244485810399 Epoch: 57 -> Test Accuracy: 84.12 [58, 60] loss: 0.198 [58, 120] loss: 0.187 [58, 180] loss: 0.208 [58, 240] loss: 0.220 [58, 300] loss: 0.204 [58, 360] loss: 0.211 Epoch: 58 -> Loss: 0.179825395346 Epoch: 58 -> Test Accuracy: 84.47 [59, 60] loss: 0.177 [59, 120] loss: 0.195 [59, 180] loss: 0.203 [59, 240] loss: 0.196 [59, 300] loss: 0.193 [59, 360] loss: 0.221 Epoch: 59 -> Loss: 0.152445226908 Epoch: 59 -> Test Accuracy: 84.19 [60, 60] loss: 0.184 [60, 120] loss: 0.200 [60, 180] loss: 0.193 [60, 240] loss: 0.190 [60, 300] loss: 0.200 [60, 360] loss: 0.204 Epoch: 60 -> Loss: 0.235208660364 Epoch: 60 -> Test Accuracy: 84.67 [61, 60] loss: 0.188 [61, 120] loss: 0.180 [61, 180] loss: 0.198 [61, 240] loss: 0.188 [61, 300] loss: 0.205 [61, 360] loss: 0.198 Epoch: 61 -> Loss: 0.451281130314 Epoch: 61 -> Test Accuracy: 84.23 [62, 60] loss: 0.175 [62, 120] loss: 0.174 [62, 180] loss: 0.188 [62, 240] loss: 0.189 [62, 300] loss: 0.196 [62, 360] loss: 0.205 Epoch: 62 -> Loss: 0.247376725078 Epoch: 62 -> Test Accuracy: 84.43 [63, 60] loss: 0.196 [63, 120] loss: 0.197 [63, 180] loss: 0.186 [63, 240] loss: 0.194 [63, 300] loss: 0.200 [63, 360] loss: 0.202 Epoch: 63 -> Loss: 0.269571125507 Epoch: 63 -> Test Accuracy: 84.12 [64, 60] loss: 0.182 [64, 120] loss: 0.193 [64, 180] loss: 0.191 [64, 240] loss: 0.192 [64, 300] loss: 0.200 [64, 360] loss: 0.208 Epoch: 64 -> Loss: 0.206095099449 Epoch: 64 -> Test Accuracy: 84.52 [65, 60] loss: 0.176 [65, 120] loss: 0.176 [65, 180] loss: 0.189 [65, 240] loss: 0.208 [65, 300] loss: 0.196 [65, 360] loss: 0.201 Epoch: 65 -> Loss: 0.158252879977 Epoch: 65 -> Test Accuracy: 84.07 [66, 60] loss: 0.177 [66, 120] loss: 0.187 [66, 180] loss: 0.190 [66, 240] loss: 0.193 [66, 300] loss: 0.185 [66, 360] loss: 0.196 Epoch: 66 -> Loss: 0.140484184027 Epoch: 66 -> Test Accuracy: 84.02 [67, 60] loss: 0.170 [67, 120] loss: 0.174 [67, 180] loss: 0.188 [67, 240] loss: 0.186 [67, 300] loss: 0.189 [67, 360] loss: 0.201 Epoch: 67 -> Loss: 0.203672528267 Epoch: 67 -> Test Accuracy: 84.22 [68, 60] loss: 0.181 [68, 120] loss: 0.179 [68, 180] loss: 0.185 [68, 240] loss: 0.201 [68, 300] loss: 0.194 [68, 360] loss: 0.206 Epoch: 68 -> Loss: 0.202317878604 Epoch: 68 -> Test Accuracy: 84.31 [69, 60] loss: 0.183 [69, 120] loss: 0.176 [69, 180] loss: 0.183 [69, 240] loss: 0.189 [69, 300] loss: 0.201 [69, 360] loss: 0.200 Epoch: 69 -> Loss: 0.284465789795 Epoch: 69 -> Test Accuracy: 84.04 [70, 60] loss: 0.180 [70, 120] loss: 0.195 [70, 180] loss: 0.192 [70, 240] loss: 0.194 [70, 300] loss: 0.196 [70, 360] loss: 0.200 Epoch: 70 -> Loss: 0.233686923981 Epoch: 70 -> Test Accuracy: 83.19 [71, 60] loss: 0.161 [71, 120] loss: 0.140 [71, 180] loss: 0.143 [71, 240] loss: 0.137 [71, 300] loss: 0.127 [71, 360] loss: 0.134 Epoch: 71 -> Loss: 0.0922329425812 Epoch: 71 -> Test Accuracy: 85.51 [72, 60] loss: 0.123 [72, 120] loss: 0.124 [72, 180] loss: 0.128 [72, 240] loss: 0.121 [72, 300] loss: 0.132 [72, 360] loss: 0.133 Epoch: 72 -> Loss: 0.224260404706 Epoch: 72 -> Test Accuracy: 85.19 [73, 60] loss: 0.120 [73, 120] loss: 0.116 [73, 180] loss: 0.117 [73, 240] loss: 0.123 [73, 300] loss: 0.125 [73, 360] loss: 0.117 Epoch: 73 -> Loss: 0.129956111312 Epoch: 73 -> Test Accuracy: 85.27 [74, 60] loss: 0.103 [74, 120] loss: 0.106 [74, 180] loss: 0.114 [74, 240] loss: 0.120 [74, 300] loss: 0.115 [74, 360] loss: 0.119 Epoch: 74 -> Loss: 0.0741558670998 Epoch: 74 -> Test Accuracy: 85.06 [75, 60] loss: 0.110 [75, 120] loss: 0.103 [75, 180] loss: 0.113 [75, 240] loss: 0.103 [75, 300] loss: 0.106 [75, 360] loss: 0.114 Epoch: 75 -> Loss: 0.144316285849 Epoch: 75 -> Test Accuracy: 85.28 [76, 60] loss: 0.106 [76, 120] loss: 0.105 [76, 180] loss: 0.108 [76, 240] loss: 0.107 [76, 300] loss: 0.113 [76, 360] loss: 0.106 Epoch: 76 -> Loss: 0.132021829486 Epoch: 76 -> Test Accuracy: 85.41 [77, 60] loss: 0.101 [77, 120] loss: 0.115 [77, 180] loss: 0.105 [77, 240] loss: 0.103 [77, 300] loss: 0.100 [77, 360] loss: 0.102 Epoch: 77 -> Loss: 0.0991534739733 Epoch: 77 -> Test Accuracy: 84.98 [78, 60] loss: 0.103 [78, 120] loss: 0.095 [78, 180] loss: 0.103 [78, 240] loss: 0.104 [78, 300] loss: 0.105 [78, 360] loss: 0.104 Epoch: 78 -> Loss: 0.146122738719 Epoch: 78 -> Test Accuracy: 85.19 [79, 60] loss: 0.095 [79, 120] loss: 0.098 [79, 180] loss: 0.107 [79, 240] loss: 0.099 [79, 300] loss: 0.096 [79, 360] loss: 0.101 Epoch: 79 -> Loss: 0.133021309972 Epoch: 79 -> Test Accuracy: 85.07 [80, 60] loss: 0.096 [80, 120] loss: 0.096 [80, 180] loss: 0.101 [80, 240] loss: 0.105 [80, 300] loss: 0.094 [80, 360] loss: 0.104 Epoch: 80 -> Loss: 0.159720659256 Epoch: 80 -> Test Accuracy: 85.03 [81, 60] loss: 0.093 [81, 120] loss: 0.094 [81, 180] loss: 0.093 [81, 240] loss: 0.098 [81, 300] loss: 0.099 [81, 360] loss: 0.099 Epoch: 81 -> Loss: 0.158382326365 Epoch: 81 -> Test Accuracy: 85.46 [82, 60] loss: 0.095 [82, 120] loss: 0.090 [82, 180] loss: 0.093 [82, 240] loss: 0.091 [82, 300] loss: 0.093 [82, 360] loss: 0.095 Epoch: 82 -> Loss: 0.0606381893158 Epoch: 82 -> Test Accuracy: 85.38 [83, 60] loss: 0.090 [83, 120] loss: 0.092 [83, 180] loss: 0.097 [83, 240] loss: 0.097 [83, 300] loss: 0.100 [83, 360] loss: 0.100 Epoch: 83 -> Loss: 0.0515924468637 Epoch: 83 -> Test Accuracy: 85.29 [84, 60] loss: 0.091 [84, 120] loss: 0.091 [84, 180] loss: 0.097 [84, 240] loss: 0.090 [84, 300] loss: 0.094 [84, 360] loss: 0.095 Epoch: 84 -> Loss: 0.0481597967446 Epoch: 84 -> Test Accuracy: 85.14 [85, 60] loss: 0.083 [85, 120] loss: 0.092 [85, 180] loss: 0.101 [85, 240] loss: 0.089 [85, 300] loss: 0.099 [85, 360] loss: 0.085 Epoch: 85 -> Loss: 0.0471388734877 Epoch: 85 -> Test Accuracy: 85.06 [86, 60] loss: 0.082 [86, 120] loss: 0.081 [86, 180] loss: 0.083 [86, 240] loss: 0.084 [86, 300] loss: 0.081 [86, 360] loss: 0.079 Epoch: 86 -> Loss: 0.0917103737593 Epoch: 86 -> Test Accuracy: 85.27 [87, 60] loss: 0.081 [87, 120] loss: 0.077 [87, 180] loss: 0.086 [87, 240] loss: 0.082 [87, 300] loss: 0.080 [87, 360] loss: 0.081 Epoch: 87 -> Loss: 0.109203979373 Epoch: 87 -> Test Accuracy: 85.25 [88, 60] loss: 0.079 [88, 120] loss: 0.084 [88, 180] loss: 0.084 [88, 240] loss: 0.078 [88, 300] loss: 0.080 [88, 360] loss: 0.080 Epoch: 88 -> Loss: 0.0930551737547 Epoch: 88 -> Test Accuracy: 85.26 [89, 60] loss: 0.080 [89, 120] loss: 0.079 [89, 180] loss: 0.078 [89, 240] loss: 0.079 [89, 300] loss: 0.077 [89, 360] loss: 0.079 Epoch: 89 -> Loss: 0.089452907443 Epoch: 89 -> Test Accuracy: 85.32 [90, 60] loss: 0.078 [90, 120] loss: 0.077 [90, 180] loss: 0.073 [90, 240] loss: 0.076 [90, 300] loss: 0.071 [90, 360] loss: 0.076 Epoch: 90 -> Loss: 0.0825238227844 Epoch: 90 -> Test Accuracy: 85.39 [91, 60] loss: 0.072 [91, 120] loss: 0.077 [91, 180] loss: 0.076 [91, 240] loss: 0.075 [91, 300] loss: 0.080 [91, 360] loss: 0.078 Epoch: 91 -> Loss: 0.154678195715 Epoch: 91 -> Test Accuracy: 85.29 [92, 60] loss: 0.074 [92, 120] loss: 0.075 [92, 180] loss: 0.076 [92, 240] loss: 0.074 [92, 300] loss: 0.082 [92, 360] loss: 0.079 Epoch: 92 -> Loss: 0.0641233474016 Epoch: 92 -> Test Accuracy: 85.36 [93, 60] loss: 0.074 [93, 120] loss: 0.078 [93, 180] loss: 0.072 [93, 240] loss: 0.078 [93, 300] loss: 0.072 [93, 360] loss: 0.075 Epoch: 93 -> Loss: 0.0602846257389 Epoch: 93 -> Test Accuracy: 85.43 [94, 60] loss: 0.076 [94, 120] loss: 0.079 [94, 180] loss: 0.076 [94, 240] loss: 0.076 [94, 300] loss: 0.074 [94, 360] loss: 0.072 Epoch: 94 -> Loss: 0.0833013057709 Epoch: 94 -> Test Accuracy: 85.43 [95, 60] loss: 0.078 [95, 120] loss: 0.077 [95, 180] loss: 0.077 [95, 240] loss: 0.073 [95, 300] loss: 0.072 [95, 360] loss: 0.074 Epoch: 95 -> Loss: 0.0787304118276 Epoch: 95 -> Test Accuracy: 85.39 [96, 60] loss: 0.078 [96, 120] loss: 0.072 [96, 180] loss: 0.076 [96, 240] loss: 0.069 [96, 300] loss: 0.071 [96, 360] loss: 0.075 Epoch: 96 -> Loss: 0.0581420250237 Epoch: 96 -> Test Accuracy: 85.25 [97, 60] loss: 0.072 [97, 120] loss: 0.077 [97, 180] loss: 0.073 [97, 240] loss: 0.078 [97, 300] loss: 0.069 [97, 360] loss: 0.072 Epoch: 97 -> Loss: 0.0873824954033 Epoch: 97 -> Test Accuracy: 85.24 [98, 60] loss: 0.076 [98, 120] loss: 0.073 [98, 180] loss: 0.070 [98, 240] loss: 0.072 [98, 300] loss: 0.071 [98, 360] loss: 0.077 Epoch: 98 -> Loss: 0.108136489987 Epoch: 98 -> Test Accuracy: 85.39 [99, 60] loss: 0.071 [99, 120] loss: 0.069 [99, 180] loss: 0.073 [99, 240] loss: 0.071 [99, 300] loss: 0.074 [99, 360] loss: 0.071 Epoch: 99 -> Loss: 0.0875692218542 Epoch: 99 -> Test Accuracy: 85.34 [100, 60] loss: 0.070 [100, 120] loss: 0.075 [100, 180] loss: 0.065 [100, 240] loss: 0.072 [100, 300] loss: 0.075 [100, 360] loss: 0.077 Epoch: 100 -> Loss: 0.106583692133 Epoch: 100 -> Test Accuracy: 85.4 Finished Training [1, 60] loss: 1.230 [1, 120] loss: 0.980 [1, 180] loss: 0.939 [1, 240] loss: 0.893 [1, 300] loss: 0.863 [1, 360] loss: 0.857 Epoch: 1 -> Loss: 0.650463223457 Epoch: 1 -> Test Accuracy: 65.72 [2, 60] loss: 0.830 [2, 120] loss: 0.812 [2, 180] loss: 0.803 [2, 240] loss: 0.768 [2, 300] loss: 0.772 [2, 360] loss: 0.765 Epoch: 2 -> Loss: 0.787874281406 Epoch: 2 -> Test Accuracy: 66.46 [3, 60] loss: 0.755 [3, 120] loss: 0.732 [3, 180] loss: 0.755 [3, 240] loss: 0.727 [3, 300] loss: 0.750 [3, 360] loss: 0.732 Epoch: 3 -> Loss: 0.760669112206 Epoch: 3 -> Test Accuracy: 68.46 [4, 60] loss: 0.730 [4, 120] loss: 0.719 [4, 180] loss: 0.743 [4, 240] loss: 0.722 [4, 300] loss: 0.713 [4, 360] loss: 0.717 Epoch: 4 -> Loss: 0.802625060081 Epoch: 4 -> Test Accuracy: 68.2 [5, 60] loss: 0.711 [5, 120] loss: 0.716 [5, 180] loss: 0.717 [5, 240] loss: 0.729 [5, 300] loss: 0.716 [5, 360] loss: 0.698 Epoch: 5 -> Loss: 0.886916339397 Epoch: 5 -> Test Accuracy: 69.89 [6, 60] loss: 0.679 [6, 120] loss: 0.713 [6, 180] loss: 0.700 [6, 240] loss: 0.700 [6, 300] loss: 0.697 [6, 360] loss: 0.688 Epoch: 6 -> Loss: 0.740843892097 Epoch: 6 -> Test Accuracy: 68.83 [7, 60] loss: 0.683 [7, 120] loss: 0.692 [7, 180] loss: 0.696 [7, 240] loss: 0.700 [7, 300] loss: 0.688 [7, 360] loss: 0.713 Epoch: 7 -> Loss: 0.778291881084 Epoch: 7 -> Test Accuracy: 70.97 [8, 60] loss: 0.684 [8, 120] loss: 0.664 [8, 180] loss: 0.707 [8, 240] loss: 0.689 [8, 300] loss: 0.683 [8, 360] loss: 0.702 Epoch: 8 -> Loss: 0.677375733852 Epoch: 8 -> Test Accuracy: 69.92 [9, 60] loss: 0.672 [9, 120] loss: 0.660 [9, 180] loss: 0.681 [9, 240] loss: 0.663 [9, 300] loss: 0.698 [9, 360] loss: 0.689 Epoch: 9 -> Loss: 0.838647007942 Epoch: 9 -> Test Accuracy: 68.99 [10, 60] loss: 0.657 [10, 120] loss: 0.669 [10, 180] loss: 0.682 [10, 240] loss: 0.682 [10, 300] loss: 0.686 [10, 360] loss: 0.682 Epoch: 10 -> Loss: 0.778439223766 Epoch: 10 -> Test Accuracy: 69.34 [11, 60] loss: 0.657 [11, 120] loss: 0.676 [11, 180] loss: 0.665 [11, 240] loss: 0.688 [11, 300] loss: 0.667 [11, 360] loss: 0.690 Epoch: 11 -> Loss: 0.775718033314 Epoch: 11 -> Test Accuracy: 70.35 [12, 60] loss: 0.660 [12, 120] loss: 0.679 [12, 180] loss: 0.658 [12, 240] loss: 0.700 [12, 300] loss: 0.669 [12, 360] loss: 0.687 Epoch: 12 -> Loss: 0.702473163605 Epoch: 12 -> Test Accuracy: 70.93 [13, 60] loss: 0.659 [13, 120] loss: 0.671 [13, 180] loss: 0.657 [13, 240] loss: 0.670 [13, 300] loss: 0.669 [13, 360] loss: 0.670 Epoch: 13 -> Loss: 0.753682732582 Epoch: 13 -> Test Accuracy: 70.62 [14, 60] loss: 0.638 [14, 120] loss: 0.675 [14, 180] loss: 0.654 [14, 240] loss: 0.655 [14, 300] loss: 0.687 [14, 360] loss: 0.666 Epoch: 14 -> Loss: 0.894064247608 Epoch: 14 -> Test Accuracy: 69.9 [15, 60] loss: 0.645 [15, 120] loss: 0.653 [15, 180] loss: 0.676 [15, 240] loss: 0.659 [15, 300] loss: 0.663 [15, 360] loss: 0.663 Epoch: 15 -> Loss: 0.635680139065 Epoch: 15 -> Test Accuracy: 69.58 [16, 60] loss: 0.651 [16, 120] loss: 0.651 [16, 180] loss: 0.672 [16, 240] loss: 0.658 [16, 300] loss: 0.680 [16, 360] loss: 0.669 Epoch: 16 -> Loss: 0.502421319485 Epoch: 16 -> Test Accuracy: 70.04 [17, 60] loss: 0.660 [17, 120] loss: 0.648 [17, 180] loss: 0.658 [17, 240] loss: 0.648 [17, 300] loss: 0.649 [17, 360] loss: 0.671 Epoch: 17 -> Loss: 0.704840838909 Epoch: 17 -> Test Accuracy: 70.54 [18, 60] loss: 0.644 [18, 120] loss: 0.661 [18, 180] loss: 0.676 [18, 240] loss: 0.688 [18, 300] loss: 0.675 [18, 360] loss: 0.652 Epoch: 18 -> Loss: 0.815052688122 Epoch: 18 -> Test Accuracy: 70.73 [19, 60] loss: 0.649 [19, 120] loss: 0.635 [19, 180] loss: 0.679 [19, 240] loss: 0.674 [19, 300] loss: 0.653 [19, 360] loss: 0.660 Epoch: 19 -> Loss: 0.764343142509 Epoch: 19 -> Test Accuracy: 70.64 [20, 60] loss: 0.663 [20, 120] loss: 0.650 [20, 180] loss: 0.639 [20, 240] loss: 0.665 [20, 300] loss: 0.666 [20, 360] loss: 0.654 Epoch: 20 -> Loss: 0.586040139198 Epoch: 20 -> Test Accuracy: 69.56 [21, 60] loss: 0.656 [21, 120] loss: 0.645 [21, 180] loss: 0.643 [21, 240] loss: 0.676 [21, 300] loss: 0.637 [21, 360] loss: 0.654 Epoch: 21 -> Loss: 0.644037604332 Epoch: 21 -> Test Accuracy: 71.2 [22, 60] loss: 0.651 [22, 120] loss: 0.657 [22, 180] loss: 0.653 [22, 240] loss: 0.647 [22, 300] loss: 0.653 [22, 360] loss: 0.662 Epoch: 22 -> Loss: 0.812040984631 Epoch: 22 -> Test Accuracy: 71.79 [23, 60] loss: 0.658 [23, 120] loss: 0.624 [23, 180] loss: 0.651 [23, 240] loss: 0.638 [23, 300] loss: 0.644 [23, 360] loss: 0.677 Epoch: 23 -> Loss: 0.578508019447 Epoch: 23 -> Test Accuracy: 71.05 [24, 60] loss: 0.635 [24, 120] loss: 0.646 [24, 180] loss: 0.664 [24, 240] loss: 0.664 [24, 300] loss: 0.655 [24, 360] loss: 0.666 Epoch: 24 -> Loss: 0.662880301476 Epoch: 24 -> Test Accuracy: 71.53 [25, 60] loss: 0.630 [25, 120] loss: 0.651 [25, 180] loss: 0.647 [25, 240] loss: 0.658 [25, 300] loss: 0.638 [25, 360] loss: 0.661 Epoch: 25 -> Loss: 0.739185631275 Epoch: 25 -> Test Accuracy: 70.33 [26, 60] loss: 0.669 [26, 120] loss: 0.654 [26, 180] loss: 0.658 [26, 240] loss: 0.659 [26, 300] loss: 0.636 [26, 360] loss: 0.661 Epoch: 26 -> Loss: 0.637153983116 Epoch: 26 -> Test Accuracy: 70.77 [27, 60] loss: 0.633 [27, 120] loss: 0.658 [27, 180] loss: 0.650 [27, 240] loss: 0.650 [27, 300] loss: 0.646 [27, 360] loss: 0.659 Epoch: 27 -> Loss: 0.662607073784 Epoch: 27 -> Test Accuracy: 71.39 [28, 60] loss: 0.655 [28, 120] loss: 0.652 [28, 180] loss: 0.639 [28, 240] loss: 0.645 [28, 300] loss: 0.654 [28, 360] loss: 0.661 Epoch: 28 -> Loss: 0.467119842768 Epoch: 28 -> Test Accuracy: 72.14 [29, 60] loss: 0.635 [29, 120] loss: 0.631 [29, 180] loss: 0.654 [29, 240] loss: 0.666 [29, 300] loss: 0.639 [29, 360] loss: 0.650 Epoch: 29 -> Loss: 0.704536378384 Epoch: 29 -> Test Accuracy: 69.67 [30, 60] loss: 0.662 [30, 120] loss: 0.636 [30, 180] loss: 0.637 [30, 240] loss: 0.650 [30, 300] loss: 0.651 [30, 360] loss: 0.668 Epoch: 30 -> Loss: 0.742483079433 Epoch: 30 -> Test Accuracy: 71.63 [31, 60] loss: 0.655 [31, 120] loss: 0.605 [31, 180] loss: 0.647 [31, 240] loss: 0.679 [31, 300] loss: 0.639 [31, 360] loss: 0.677 Epoch: 31 -> Loss: 0.812393546104 Epoch: 31 -> Test Accuracy: 70.79 [32, 60] loss: 0.630 [32, 120] loss: 0.649 [32, 180] loss: 0.675 [32, 240] loss: 0.629 [32, 300] loss: 0.654 [32, 360] loss: 0.659 Epoch: 32 -> Loss: 0.834937930107 Epoch: 32 -> Test Accuracy: 70.19 [33, 60] loss: 0.632 [33, 120] loss: 0.665 [33, 180] loss: 0.652 [33, 240] loss: 0.645 [33, 300] loss: 0.647 [33, 360] loss: 0.658 Epoch: 33 -> Loss: 0.631586313248 Epoch: 33 -> Test Accuracy: 71.43 [34, 60] loss: 0.632 [34, 120] loss: 0.657 [34, 180] loss: 0.656 [34, 240] loss: 0.646 [34, 300] loss: 0.636 [34, 360] loss: 0.644 Epoch: 34 -> Loss: 0.700482487679 Epoch: 34 -> Test Accuracy: 70.64 [35, 60] loss: 0.639 [35, 120] loss: 0.652 [35, 180] loss: 0.658 [35, 240] loss: 0.633 [35, 300] loss: 0.635 [35, 360] loss: 0.643 Epoch: 35 -> Loss: 0.765577435493 Epoch: 35 -> Test Accuracy: 70.54 [36, 60] loss: 0.598 [36, 120] loss: 0.574 [36, 180] loss: 0.555 [36, 240] loss: 0.571 [36, 300] loss: 0.544 [36, 360] loss: 0.560 Epoch: 36 -> Loss: 0.447709709406 Epoch: 36 -> Test Accuracy: 74.07 [37, 60] loss: 0.555 [37, 120] loss: 0.527 [37, 180] loss: 0.550 [37, 240] loss: 0.546 [37, 300] loss: 0.532 [37, 360] loss: 0.538 Epoch: 37 -> Loss: 0.494886398315 Epoch: 37 -> Test Accuracy: 74.87 [38, 60] loss: 0.531 [38, 120] loss: 0.521 [38, 180] loss: 0.538 [38, 240] loss: 0.539 [38, 300] loss: 0.534 [38, 360] loss: 0.549 Epoch: 38 -> Loss: 0.617129087448 Epoch: 38 -> Test Accuracy: 74.88 [39, 60] loss: 0.527 [39, 120] loss: 0.535 [39, 180] loss: 0.503 [39, 240] loss: 0.530 [39, 300] loss: 0.529 [39, 360] loss: 0.533 Epoch: 39 -> Loss: 0.571642994881 Epoch: 39 -> Test Accuracy: 74.93 [40, 60] loss: 0.505 [40, 120] loss: 0.526 [40, 180] loss: 0.524 [40, 240] loss: 0.539 [40, 300] loss: 0.523 [40, 360] loss: 0.529 Epoch: 40 -> Loss: 0.464881032705 Epoch: 40 -> Test Accuracy: 74.44 [41, 60] loss: 0.526 [41, 120] loss: 0.514 [41, 180] loss: 0.512 [41, 240] loss: 0.526 [41, 300] loss: 0.508 [41, 360] loss: 0.529 Epoch: 41 -> Loss: 0.545656323433 Epoch: 41 -> Test Accuracy: 75.06 [42, 60] loss: 0.517 [42, 120] loss: 0.541 [42, 180] loss: 0.517 [42, 240] loss: 0.511 [42, 300] loss: 0.530 [42, 360] loss: 0.529 Epoch: 42 -> Loss: 0.519388616085 Epoch: 42 -> Test Accuracy: 74.73 [43, 60] loss: 0.511 [43, 120] loss: 0.527 [43, 180] loss: 0.495 [43, 240] loss: 0.513 [43, 300] loss: 0.539 [43, 360] loss: 0.530 Epoch: 43 -> Loss: 0.384705603123 Epoch: 43 -> Test Accuracy: 74.91 [44, 60] loss: 0.510 [44, 120] loss: 0.501 [44, 180] loss: 0.527 [44, 240] loss: 0.525 [44, 300] loss: 0.526 [44, 360] loss: 0.534 Epoch: 44 -> Loss: 0.535769581795 Epoch: 44 -> Test Accuracy: 74.7 [45, 60] loss: 0.512 [45, 120] loss: 0.514 [45, 180] loss: 0.518 [45, 240] loss: 0.537 [45, 300] loss: 0.518 [45, 360] loss: 0.531 Epoch: 45 -> Loss: 0.515571177006 Epoch: 45 -> Test Accuracy: 74.67 [46, 60] loss: 0.508 [46, 120] loss: 0.496 [46, 180] loss: 0.519 [46, 240] loss: 0.521 [46, 300] loss: 0.527 [46, 360] loss: 0.548 Epoch: 46 -> Loss: 0.589191794395 Epoch: 46 -> Test Accuracy: 74.65 [47, 60] loss: 0.512 [47, 120] loss: 0.509 [47, 180] loss: 0.522 [47, 240] loss: 0.511 [47, 300] loss: 0.526 [47, 360] loss: 0.536 Epoch: 47 -> Loss: 0.578607082367 Epoch: 47 -> Test Accuracy: 74.46 [48, 60] loss: 0.521 [48, 120] loss: 0.514 [48, 180] loss: 0.534 [48, 240] loss: 0.491 [48, 300] loss: 0.517 [48, 360] loss: 0.530 Epoch: 48 -> Loss: 0.517775058746 Epoch: 48 -> Test Accuracy: 74.73 [49, 60] loss: 0.505 [49, 120] loss: 0.512 [49, 180] loss: 0.503 [49, 240] loss: 0.529 [49, 300] loss: 0.526 [49, 360] loss: 0.511 Epoch: 49 -> Loss: 0.652405083179 Epoch: 49 -> Test Accuracy: 73.27 [50, 60] loss: 0.507 [50, 120] loss: 0.517 [50, 180] loss: 0.520 [50, 240] loss: 0.529 [50, 300] loss: 0.529 [50, 360] loss: 0.522 Epoch: 50 -> Loss: 0.509332418442 Epoch: 50 -> Test Accuracy: 74.12 [51, 60] loss: 0.525 [51, 120] loss: 0.522 [51, 180] loss: 0.502 [51, 240] loss: 0.521 [51, 300] loss: 0.523 [51, 360] loss: 0.530 Epoch: 51 -> Loss: 0.664568245411 Epoch: 51 -> Test Accuracy: 73.92 [52, 60] loss: 0.502 [52, 120] loss: 0.504 [52, 180] loss: 0.503 [52, 240] loss: 0.524 [52, 300] loss: 0.538 [52, 360] loss: 0.502 Epoch: 52 -> Loss: 0.807490944862 Epoch: 52 -> Test Accuracy: 74.28 [53, 60] loss: 0.506 [53, 120] loss: 0.504 [53, 180] loss: 0.532 [53, 240] loss: 0.544 [53, 300] loss: 0.526 [53, 360] loss: 0.513 Epoch: 53 -> Loss: 0.50402867794 Epoch: 53 -> Test Accuracy: 74.45 [54, 60] loss: 0.510 [54, 120] loss: 0.498 [54, 180] loss: 0.525 [54, 240] loss: 0.520 [54, 300] loss: 0.522 [54, 360] loss: 0.513 Epoch: 54 -> Loss: 0.543969511986 Epoch: 54 -> Test Accuracy: 74.06 [55, 60] loss: 0.516 [55, 120] loss: 0.514 [55, 180] loss: 0.541 [55, 240] loss: 0.508 [55, 300] loss: 0.524 [55, 360] loss: 0.534 Epoch: 55 -> Loss: 0.437804788351 Epoch: 55 -> Test Accuracy: 74.47 [56, 60] loss: 0.510 [56, 120] loss: 0.505 [56, 180] loss: 0.513 [56, 240] loss: 0.513 [56, 300] loss: 0.513 [56, 360] loss: 0.529 Epoch: 56 -> Loss: 0.575148999691 Epoch: 56 -> Test Accuracy: 74.21 [57, 60] loss: 0.505 [57, 120] loss: 0.515 [57, 180] loss: 0.527 [57, 240] loss: 0.499 [57, 300] loss: 0.516 [57, 360] loss: 0.529 Epoch: 57 -> Loss: 0.565602242947 Epoch: 57 -> Test Accuracy: 74.27 [58, 60] loss: 0.508 [58, 120] loss: 0.498 [58, 180] loss: 0.513 [58, 240] loss: 0.524 [58, 300] loss: 0.529 [58, 360] loss: 0.522 Epoch: 58 -> Loss: 0.326432406902 Epoch: 58 -> Test Accuracy: 74.3 [59, 60] loss: 0.490 [59, 120] loss: 0.527 [59, 180] loss: 0.512 [59, 240] loss: 0.501 [59, 300] loss: 0.526 [59, 360] loss: 0.512 Epoch: 59 -> Loss: 0.586781084538 Epoch: 59 -> Test Accuracy: 74.1 [60, 60] loss: 0.513 [60, 120] loss: 0.529 [60, 180] loss: 0.522 [60, 240] loss: 0.516 [60, 300] loss: 0.524 [60, 360] loss: 0.517 Epoch: 60 -> Loss: 0.510123133659 Epoch: 60 -> Test Accuracy: 74.74 [61, 60] loss: 0.502 [61, 120] loss: 0.503 [61, 180] loss: 0.502 [61, 240] loss: 0.528 [61, 300] loss: 0.507 [61, 360] loss: 0.510 Epoch: 61 -> Loss: 0.695623755455 Epoch: 61 -> Test Accuracy: 74.29 [62, 60] loss: 0.520 [62, 120] loss: 0.508 [62, 180] loss: 0.500 [62, 240] loss: 0.528 [62, 300] loss: 0.488 [62, 360] loss: 0.528 Epoch: 62 -> Loss: 0.405204713345 Epoch: 62 -> Test Accuracy: 74.26 [63, 60] loss: 0.507 [63, 120] loss: 0.519 [63, 180] loss: 0.518 [63, 240] loss: 0.499 [63, 300] loss: 0.488 [63, 360] loss: 0.515 Epoch: 63 -> Loss: 0.57899081707 Epoch: 63 -> Test Accuracy: 74.48 [64, 60] loss: 0.482 [64, 120] loss: 0.515 [64, 180] loss: 0.505 [64, 240] loss: 0.511 [64, 300] loss: 0.518 [64, 360] loss: 0.519 Epoch: 64 -> Loss: 0.512594163418 Epoch: 64 -> Test Accuracy: 74.79 [65, 60] loss: 0.497 [65, 120] loss: 0.500 [65, 180] loss: 0.516 [65, 240] loss: 0.497 [65, 300] loss: 0.523 [65, 360] loss: 0.525 Epoch: 65 -> Loss: 0.457039773464 Epoch: 65 -> Test Accuracy: 74.52 [66, 60] loss: 0.512 [66, 120] loss: 0.508 [66, 180] loss: 0.500 [66, 240] loss: 0.509 [66, 300] loss: 0.511 [66, 360] loss: 0.513 Epoch: 66 -> Loss: 0.532711863518 Epoch: 66 -> Test Accuracy: 74.92 [67, 60] loss: 0.491 [67, 120] loss: 0.502 [67, 180] loss: 0.518 [67, 240] loss: 0.533 [67, 300] loss: 0.503 [67, 360] loss: 0.503 Epoch: 67 -> Loss: 0.493550598621 Epoch: 67 -> Test Accuracy: 74.01 [68, 60] loss: 0.505 [68, 120] loss: 0.493 [68, 180] loss: 0.529 [68, 240] loss: 0.501 [68, 300] loss: 0.504 [68, 360] loss: 0.534 Epoch: 68 -> Loss: 0.395356237888 Epoch: 68 -> Test Accuracy: 73.78 [69, 60] loss: 0.495 [69, 120] loss: 0.506 [69, 180] loss: 0.498 [69, 240] loss: 0.506 [69, 300] loss: 0.522 [69, 360] loss: 0.502 Epoch: 69 -> Loss: 0.556714773178 Epoch: 69 -> Test Accuracy: 73.74 [70, 60] loss: 0.503 [70, 120] loss: 0.500 [70, 180] loss: 0.513 [70, 240] loss: 0.516 [70, 300] loss: 0.500 [70, 360] loss: 0.507 Epoch: 70 -> Loss: 0.507016658783 Epoch: 70 -> Test Accuracy: 74.49 [71, 60] loss: 0.471 [71, 120] loss: 0.452 [71, 180] loss: 0.427 [71, 240] loss: 0.449 [71, 300] loss: 0.440 [71, 360] loss: 0.470 Epoch: 71 -> Loss: 0.573578238487 Epoch: 71 -> Test Accuracy: 76.35 [72, 60] loss: 0.448 [72, 120] loss: 0.426 [72, 180] loss: 0.451 [72, 240] loss: 0.441 [72, 300] loss: 0.438 [72, 360] loss: 0.428 Epoch: 72 -> Loss: 0.490967571735 Epoch: 72 -> Test Accuracy: 76.48 [73, 60] loss: 0.424 [73, 120] loss: 0.436 [73, 180] loss: 0.429 [73, 240] loss: 0.442 [73, 300] loss: 0.442 [73, 360] loss: 0.426 Epoch: 73 -> Loss: 0.503980875015 Epoch: 73 -> Test Accuracy: 76.69 [74, 60] loss: 0.429 [74, 120] loss: 0.425 [74, 180] loss: 0.429 [74, 240] loss: 0.423 [74, 300] loss: 0.422 [74, 360] loss: 0.415 Epoch: 74 -> Loss: 0.602517187595 Epoch: 74 -> Test Accuracy: 76.84 [75, 60] loss: 0.409 [75, 120] loss: 0.426 [75, 180] loss: 0.433 [75, 240] loss: 0.426 [75, 300] loss: 0.414 [75, 360] loss: 0.439 Epoch: 75 -> Loss: 0.393859446049 Epoch: 75 -> Test Accuracy: 76.43 [76, 60] loss: 0.413 [76, 120] loss: 0.426 [76, 180] loss: 0.425 [76, 240] loss: 0.430 [76, 300] loss: 0.418 [76, 360] loss: 0.417 Epoch: 76 -> Loss: 0.406775295734 Epoch: 76 -> Test Accuracy: 76.52 [77, 60] loss: 0.404 [77, 120] loss: 0.419 [77, 180] loss: 0.424 [77, 240] loss: 0.421 [77, 300] loss: 0.423 [77, 360] loss: 0.422 Epoch: 77 -> Loss: 0.478052318096 Epoch: 77 -> Test Accuracy: 76.52 [78, 60] loss: 0.413 [78, 120] loss: 0.420 [78, 180] loss: 0.412 [78, 240] loss: 0.411 [78, 300] loss: 0.409 [78, 360] loss: 0.429 Epoch: 78 -> Loss: 0.368268817663 Epoch: 78 -> Test Accuracy: 76.57 [79, 60] loss: 0.419 [79, 120] loss: 0.411 [79, 180] loss: 0.408 [79, 240] loss: 0.428 [79, 300] loss: 0.416 [79, 360] loss: 0.409 Epoch: 79 -> Loss: 0.373751223087 Epoch: 79 -> Test Accuracy: 76.82 [80, 60] loss: 0.390 [80, 120] loss: 0.410 [80, 180] loss: 0.406 [80, 240] loss: 0.410 [80, 300] loss: 0.415 [80, 360] loss: 0.420 Epoch: 80 -> Loss: 0.37625131011 Epoch: 80 -> Test Accuracy: 76.64 [81, 60] loss: 0.407 [81, 120] loss: 0.418 [81, 180] loss: 0.409 [81, 240] loss: 0.402 [81, 300] loss: 0.420 [81, 360] loss: 0.418 Epoch: 81 -> Loss: 0.494882762432 Epoch: 81 -> Test Accuracy: 76.25 [82, 60] loss: 0.410 [82, 120] loss: 0.412 [82, 180] loss: 0.407 [82, 240] loss: 0.413 [82, 300] loss: 0.416 [82, 360] loss: 0.402 Epoch: 82 -> Loss: 0.318713784218 Epoch: 82 -> Test Accuracy: 76.53 [83, 60] loss: 0.402 [83, 120] loss: 0.408 [83, 180] loss: 0.402 [83, 240] loss: 0.418 [83, 300] loss: 0.404 [83, 360] loss: 0.415 Epoch: 83 -> Loss: 0.441703498363 Epoch: 83 -> Test Accuracy: 76.49 [84, 60] loss: 0.410 [84, 120] loss: 0.409 [84, 180] loss: 0.395 [84, 240] loss: 0.411 [84, 300] loss: 0.409 [84, 360] loss: 0.405 Epoch: 84 -> Loss: 0.360123574734 Epoch: 84 -> Test Accuracy: 76.15 [85, 60] loss: 0.398 [85, 120] loss: 0.408 [85, 180] loss: 0.405 [85, 240] loss: 0.406 [85, 300] loss: 0.410 [85, 360] loss: 0.420 Epoch: 85 -> Loss: 0.449885308743 Epoch: 85 -> Test Accuracy: 76.36 [86, 60] loss: 0.393 [86, 120] loss: 0.395 [86, 180] loss: 0.381 [86, 240] loss: 0.378 [86, 300] loss: 0.375 [86, 360] loss: 0.394 Epoch: 86 -> Loss: 0.360383838415 Epoch: 86 -> Test Accuracy: 76.94 [87, 60] loss: 0.383 [87, 120] loss: 0.393 [87, 180] loss: 0.376 [87, 240] loss: 0.378 [87, 300] loss: 0.367 [87, 360] loss: 0.393 Epoch: 87 -> Loss: 0.313582003117 Epoch: 87 -> Test Accuracy: 76.93 [88, 60] loss: 0.375 [88, 120] loss: 0.373 [88, 180] loss: 0.389 [88, 240] loss: 0.381 [88, 300] loss: 0.376 [88, 360] loss: 0.380 Epoch: 88 -> Loss: 0.419721186161 Epoch: 88 -> Test Accuracy: 77.0 [89, 60] loss: 0.373 [89, 120] loss: 0.379 [89, 180] loss: 0.373 [89, 240] loss: 0.369 [89, 300] loss: 0.378 [89, 360] loss: 0.390 Epoch: 89 -> Loss: 0.305138349533 Epoch: 89 -> Test Accuracy: 77.05 [90, 60] loss: 0.376 [90, 120] loss: 0.388 [90, 180] loss: 0.370 [90, 240] loss: 0.382 [90, 300] loss: 0.379 [90, 360] loss: 0.375 Epoch: 90 -> Loss: 0.428806781769 Epoch: 90 -> Test Accuracy: 77.04 [91, 60] loss: 0.382 [91, 120] loss: 0.373 [91, 180] loss: 0.379 [91, 240] loss: 0.371 [91, 300] loss: 0.366 [91, 360] loss: 0.378 Epoch: 91 -> Loss: 0.242852404714 Epoch: 91 -> Test Accuracy: 77.2 [92, 60] loss: 0.377 [92, 120] loss: 0.363 [92, 180] loss: 0.376 [92, 240] loss: 0.370 [92, 300] loss: 0.383 [92, 360] loss: 0.380 Epoch: 92 -> Loss: 0.294145166874 Epoch: 92 -> Test Accuracy: 76.88 [93, 60] loss: 0.366 [93, 120] loss: 0.387 [93, 180] loss: 0.376 [93, 240] loss: 0.371 [93, 300] loss: 0.361 [93, 360] loss: 0.376 Epoch: 93 -> Loss: 0.481191962957 Epoch: 93 -> Test Accuracy: 77.21 [94, 60] loss: 0.363 [94, 120] loss: 0.375 [94, 180] loss: 0.370 [94, 240] loss: 0.375 [94, 300] loss: 0.385 [94, 360] loss: 0.381 Epoch: 94 -> Loss: 0.371431410313 Epoch: 94 -> Test Accuracy: 77.03 [95, 60] loss: 0.373 [95, 120] loss: 0.383 [95, 180] loss: 0.377 [95, 240] loss: 0.358 [95, 300] loss: 0.378 [95, 360] loss: 0.379 Epoch: 95 -> Loss: 0.443098962307 Epoch: 95 -> Test Accuracy: 77.05 [96, 60] loss: 0.386 [96, 120] loss: 0.369 [96, 180] loss: 0.375 [96, 240] loss: 0.371 [96, 300] loss: 0.376 [96, 360] loss: 0.370 Epoch: 96 -> Loss: 0.290161937475 Epoch: 96 -> Test Accuracy: 77.3 [97, 60] loss: 0.364 [97, 120] loss: 0.369 [97, 180] loss: 0.373 [97, 240] loss: 0.371 [97, 300] loss: 0.366 [97, 360] loss: 0.369 Epoch: 97 -> Loss: 0.397721797228 Epoch: 97 -> Test Accuracy: 77.16 [98, 60] loss: 0.369 [98, 120] loss: 0.366 [98, 180] loss: 0.364 [98, 240] loss: 0.367 [98, 300] loss: 0.370 [98, 360] loss: 0.371 Epoch: 98 -> Loss: 0.407518237829 Epoch: 98 -> Test Accuracy: 77.12 [99, 60] loss: 0.363 [99, 120] loss: 0.376 [99, 180] loss: 0.368 [99, 240] loss: 0.350 [99, 300] loss: 0.369 [99, 360] loss: 0.369 Epoch: 99 -> Loss: 0.345765531063 Epoch: 99 -> Test Accuracy: 76.94 [100, 60] loss: 0.366 [100, 120] loss: 0.351 [100, 180] loss: 0.393 [100, 240] loss: 0.368 [100, 300] loss: 0.362 [100, 360] loss: 0.379 Epoch: 100 -> Loss: 0.41921916604 Epoch: 100 -> Test Accuracy: 76.96 Finished Training [1, 60] loss: 2.214 [1, 120] loss: 2.042 [1, 180] loss: 1.986 [1, 240] loss: 1.957 [1, 300] loss: 1.930 [1, 360] loss: 1.906 Epoch: 1 -> Loss: 1.80112838745 Epoch: 1 -> Test Accuracy: 28.92 [2, 60] loss: 1.875 [2, 120] loss: 1.874 [2, 180] loss: 1.869 [2, 240] loss: 1.861 [2, 300] loss: 1.843 [2, 360] loss: 1.829 Epoch: 2 -> Loss: 1.84579432011 Epoch: 2 -> Test Accuracy: 29.05 [3, 60] loss: 1.825 [3, 120] loss: 1.802 [3, 180] loss: 1.822 [3, 240] loss: 1.818 [3, 300] loss: 1.822 [3, 360] loss: 1.799 Epoch: 3 -> Loss: 1.66067254543 Epoch: 3 -> Test Accuracy: 32.7 [4, 60] loss: 1.771 [4, 120] loss: 1.790 [4, 180] loss: 1.793 [4, 240] loss: 1.789 [4, 300] loss: 1.766 [4, 360] loss: 1.767 Epoch: 4 -> Loss: 1.83661556244 Epoch: 4 -> Test Accuracy: 31.34 [5, 60] loss: 1.773 [5, 120] loss: 1.760 [5, 180] loss: 1.760 [5, 240] loss: 1.758 [5, 300] loss: 1.768 [5, 360] loss: 1.770 Epoch: 5 -> Loss: 1.5981400013 Epoch: 5 -> Test Accuracy: 32.06 [6, 60] loss: 1.739 [6, 120] loss: 1.751 [6, 180] loss: 1.744 [6, 240] loss: 1.743 [6, 300] loss: 1.751 [6, 360] loss: 1.740 Epoch: 6 -> Loss: 1.93392252922 Epoch: 6 -> Test Accuracy: 32.84 [7, 60] loss: 1.764 [7, 120] loss: 1.720 [7, 180] loss: 1.725 [7, 240] loss: 1.734 [7, 300] loss: 1.728 [7, 360] loss: 1.743 Epoch: 7 -> Loss: 1.77970945835 Epoch: 7 -> Test Accuracy: 33.46 [8, 60] loss: 1.727 [8, 120] loss: 1.729 [8, 180] loss: 1.729 [8, 240] loss: 1.739 [8, 300] loss: 1.732 [8, 360] loss: 1.735 Epoch: 8 -> Loss: 1.62089002132 Epoch: 8 -> Test Accuracy: 32.72 [9, 60] loss: 1.725 [9, 120] loss: 1.718 [9, 180] loss: 1.719 [9, 240] loss: 1.720 [9, 300] loss: 1.718 [9, 360] loss: 1.725 Epoch: 9 -> Loss: 1.67075884342 Epoch: 9 -> Test Accuracy: 33.08 [10, 60] loss: 1.720 [10, 120] loss: 1.714 [10, 180] loss: 1.709 [10, 240] loss: 1.728 [10, 300] loss: 1.739 [10, 360] loss: 1.712 Epoch: 10 -> Loss: 1.62930130959 Epoch: 10 -> Test Accuracy: 32.78 [11, 60] loss: 1.709 [11, 120] loss: 1.707 [11, 180] loss: 1.702 [11, 240] loss: 1.704 [11, 300] loss: 1.705 [11, 360] loss: 1.709 Epoch: 11 -> Loss: 1.67310643196 Epoch: 11 -> Test Accuracy: 33.96 [12, 60] loss: 1.696 [12, 120] loss: 1.705 [12, 180] loss: 1.714 [12, 240] loss: 1.717 [12, 300] loss: 1.708 [12, 360] loss: 1.715 Epoch: 12 -> Loss: 1.63214170933 Epoch: 12 -> Test Accuracy: 34.25 [13, 60] loss: 1.707 [13, 120] loss: 1.684 [13, 180] loss: 1.707 [13, 240] loss: 1.698 [13, 300] loss: 1.696 [13, 360] loss: 1.704 Epoch: 13 -> Loss: 1.7707092762 Epoch: 13 -> Test Accuracy: 33.62 [14, 60] loss: 1.684 [14, 120] loss: 1.690 [14, 180] loss: 1.704 [14, 240] loss: 1.686 [14, 300] loss: 1.689 [14, 360] loss: 1.715 Epoch: 14 -> Loss: 1.63841688633 Epoch: 14 -> Test Accuracy: 34.7 [15, 60] loss: 1.699 [15, 120] loss: 1.684 [15, 180] loss: 1.696 [15, 240] loss: 1.691 [15, 300] loss: 1.684 [15, 360] loss: 1.699 Epoch: 15 -> Loss: 1.84138941765 Epoch: 15 -> Test Accuracy: 35.28 [16, 60] loss: 1.690 [16, 120] loss: 1.697 [16, 180] loss: 1.693 [16, 240] loss: 1.677 [16, 300] loss: 1.678 [16, 360] loss: 1.698 Epoch: 16 -> Loss: 1.51885926723 Epoch: 16 -> Test Accuracy: 35.16 [17, 60] loss: 1.692 [17, 120] loss: 1.682 [17, 180] loss: 1.704 [17, 240] loss: 1.695 [17, 300] loss: 1.677 [17, 360] loss: 1.686 Epoch: 17 -> Loss: 1.63488316536 Epoch: 17 -> Test Accuracy: 35.33 [18, 60] loss: 1.690 [18, 120] loss: 1.679 [18, 180] loss: 1.679 [18, 240] loss: 1.661 [18, 300] loss: 1.699 [18, 360] loss: 1.670 Epoch: 18 -> Loss: 1.67851006985 Epoch: 18 -> Test Accuracy: 33.34 [19, 60] loss: 1.669 [19, 120] loss: 1.687 [19, 180] loss: 1.684 [19, 240] loss: 1.695 [19, 300] loss: 1.680 [19, 360] loss: 1.696 Epoch: 19 -> Loss: 1.74293398857 Epoch: 19 -> Test Accuracy: 35.16 [20, 60] loss: 1.683 [20, 120] loss: 1.684 [20, 180] loss: 1.659 [20, 240] loss: 1.677 [20, 300] loss: 1.680 [20, 360] loss: 1.667 Epoch: 20 -> Loss: 1.69578421116 Epoch: 20 -> Test Accuracy: 34.05 [21, 60] loss: 1.661 [21, 120] loss: 1.689 [21, 180] loss: 1.681 [21, 240] loss: 1.683 [21, 300] loss: 1.697 [21, 360] loss: 1.681 Epoch: 21 -> Loss: 1.69357419014 Epoch: 21 -> Test Accuracy: 34.2 [22, 60] loss: 1.672 [22, 120] loss: 1.656 [22, 180] loss: 1.684 [22, 240] loss: 1.688 [22, 300] loss: 1.680 [22, 360] loss: 1.678 Epoch: 22 -> Loss: 1.80121481419 Epoch: 22 -> Test Accuracy: 34.95 [23, 60] loss: 1.664 [23, 120] loss: 1.688 [23, 180] loss: 1.693 [23, 240] loss: 1.673 [23, 300] loss: 1.682 [23, 360] loss: 1.679 Epoch: 23 -> Loss: 1.7194082737 Epoch: 23 -> Test Accuracy: 33.7 [24, 60] loss: 1.671 [24, 120] loss: 1.674 [24, 180] loss: 1.684 [24, 240] loss: 1.683 [24, 300] loss: 1.679 [24, 360] loss: 1.680 Epoch: 24 -> Loss: 1.72372436523 Epoch: 24 -> Test Accuracy: 33.45 [25, 60] loss: 1.678 [25, 120] loss: 1.673 [25, 180] loss: 1.665 [25, 240] loss: 1.674 [25, 300] loss: 1.673 [25, 360] loss: 1.665 Epoch: 25 -> Loss: 1.63593745232 Epoch: 25 -> Test Accuracy: 35.31 [26, 60] loss: 1.671 [26, 120] loss: 1.665 [26, 180] loss: 1.680 [26, 240] loss: 1.693 [26, 300] loss: 1.667 [26, 360] loss: 1.684 Epoch: 26 -> Loss: 1.64690184593 Epoch: 26 -> Test Accuracy: 35.33 [27, 60] loss: 1.663 [27, 120] loss: 1.686 [27, 180] loss: 1.682 [27, 240] loss: 1.666 [27, 300] loss: 1.658 [27, 360] loss: 1.670 Epoch: 27 -> Loss: 1.75495302677 Epoch: 27 -> Test Accuracy: 34.42 [28, 60] loss: 1.670 [28, 120] loss: 1.673 [28, 180] loss: 1.664 [28, 240] loss: 1.670 [28, 300] loss: 1.683 [28, 360] loss: 1.663 Epoch: 28 -> Loss: 1.61920201778 Epoch: 28 -> Test Accuracy: 35.55 [29, 60] loss: 1.672 [29, 120] loss: 1.664 [29, 180] loss: 1.645 [29, 240] loss: 1.683 [29, 300] loss: 1.665 [29, 360] loss: 1.664 Epoch: 29 -> Loss: 1.5431599617 Epoch: 29 -> Test Accuracy: 36.01 [30, 60] loss: 1.660 [30, 120] loss: 1.649 [30, 180] loss: 1.671 [30, 240] loss: 1.682 [30, 300] loss: 1.666 [30, 360] loss: 1.677 Epoch: 30 -> Loss: 1.83706569672 Epoch: 30 -> Test Accuracy: 35.54 [31, 60] loss: 1.659 [31, 120] loss: 1.652 [31, 180] loss: 1.675 [31, 240] loss: 1.680 [31, 300] loss: 1.661 [31, 360] loss: 1.663 Epoch: 31 -> Loss: 1.68681812286 Epoch: 31 -> Test Accuracy: 35.16 [32, 60] loss: 1.667 [32, 120] loss: 1.669 [32, 180] loss: 1.656 [32, 240] loss: 1.661 [32, 300] loss: 1.660 [32, 360] loss: 1.663 Epoch: 32 -> Loss: 1.65721535683 Epoch: 32 -> Test Accuracy: 35.56 [33, 60] loss: 1.656 [33, 120] loss: 1.682 [33, 180] loss: 1.672 [33, 240] loss: 1.660 [33, 300] loss: 1.668 [33, 360] loss: 1.648 Epoch: 33 -> Loss: 1.63733065128 Epoch: 33 -> Test Accuracy: 36.06 [34, 60] loss: 1.672 [34, 120] loss: 1.657 [34, 180] loss: 1.668 [34, 240] loss: 1.661 [34, 300] loss: 1.662 [34, 360] loss: 1.672 Epoch: 34 -> Loss: 1.68781399727 Epoch: 34 -> Test Accuracy: 35.21 [35, 60] loss: 1.650 [35, 120] loss: 1.643 [35, 180] loss: 1.669 [35, 240] loss: 1.657 [35, 300] loss: 1.663 [35, 360] loss: 1.669 Epoch: 35 -> Loss: 1.5745254755 Epoch: 35 -> Test Accuracy: 34.66 [36, 60] loss: 1.620 [36, 120] loss: 1.562 [36, 180] loss: 1.584 [36, 240] loss: 1.564 [36, 300] loss: 1.567 [36, 360] loss: 1.548 Epoch: 36 -> Loss: 1.50426459312 Epoch: 36 -> Test Accuracy: 39.11 [37, 60] loss: 1.558 [37, 120] loss: 1.552 [37, 180] loss: 1.524 [37, 240] loss: 1.545 [37, 300] loss: 1.557 [37, 360] loss: 1.552 Epoch: 37 -> Loss: 1.49290013313 Epoch: 37 -> Test Accuracy: 39.13 [38, 60] loss: 1.556 [38, 120] loss: 1.531 [38, 180] loss: 1.522 [38, 240] loss: 1.554 [38, 300] loss: 1.531 [38, 360] loss: 1.543 Epoch: 38 -> Loss: 1.3990226984 Epoch: 38 -> Test Accuracy: 38.42 [39, 60] loss: 1.527 [39, 120] loss: 1.546 [39, 180] loss: 1.532 [39, 240] loss: 1.543 [39, 300] loss: 1.544 [39, 360] loss: 1.526 Epoch: 39 -> Loss: 1.47474956512 Epoch: 39 -> Test Accuracy: 39.79 [40, 60] loss: 1.526 [40, 120] loss: 1.523 [40, 180] loss: 1.546 [40, 240] loss: 1.530 [40, 300] loss: 1.517 [40, 360] loss: 1.551 Epoch: 40 -> Loss: 1.53031921387 Epoch: 40 -> Test Accuracy: 38.7 [41, 60] loss: 1.540 [41, 120] loss: 1.531 [41, 180] loss: 1.545 [41, 240] loss: 1.530 [41, 300] loss: 1.533 [41, 360] loss: 1.518 Epoch: 41 -> Loss: 1.46427559853 Epoch: 41 -> Test Accuracy: 39.21 [42, 60] loss: 1.521 [42, 120] loss: 1.528 [42, 180] loss: 1.519 [42, 240] loss: 1.527 [42, 300] loss: 1.546 [42, 360] loss: 1.540 Epoch: 42 -> Loss: 1.72309815884 Epoch: 42 -> Test Accuracy: 38.96 [43, 60] loss: 1.534 [43, 120] loss: 1.545 [43, 180] loss: 1.528 [43, 240] loss: 1.529 [43, 300] loss: 1.534 [43, 360] loss: 1.532 Epoch: 43 -> Loss: 1.66689991951 Epoch: 43 -> Test Accuracy: 38.82 [44, 60] loss: 1.529 [44, 120] loss: 1.546 [44, 180] loss: 1.535 [44, 240] loss: 1.507 [44, 300] loss: 1.526 [44, 360] loss: 1.543 Epoch: 44 -> Loss: 1.5062276125 Epoch: 44 -> Test Accuracy: 38.52 [45, 60] loss: 1.530 [45, 120] loss: 1.518 [45, 180] loss: 1.550 [45, 240] loss: 1.532 [45, 300] loss: 1.552 [45, 360] loss: 1.542 Epoch: 45 -> Loss: 1.47612893581 Epoch: 45 -> Test Accuracy: 39.63 [46, 60] loss: 1.515 [46, 120] loss: 1.545 [46, 180] loss: 1.525 [46, 240] loss: 1.533 [46, 300] loss: 1.530 [46, 360] loss: 1.541 Epoch: 46 -> Loss: 1.75978207588 Epoch: 46 -> Test Accuracy: 39.52 [47, 60] loss: 1.524 [47, 120] loss: 1.533 [47, 180] loss: 1.534 [47, 240] loss: 1.552 [47, 300] loss: 1.550 [47, 360] loss: 1.523 Epoch: 47 -> Loss: 1.61171889305 Epoch: 47 -> Test Accuracy: 39.82 [48, 60] loss: 1.527 [48, 120] loss: 1.560 [48, 180] loss: 1.525 [48, 240] loss: 1.541 [48, 300] loss: 1.518 [48, 360] loss: 1.525 Epoch: 48 -> Loss: 1.39809668064 Epoch: 48 -> Test Accuracy: 38.54 [49, 60] loss: 1.521 [49, 120] loss: 1.522 [49, 180] loss: 1.534 [49, 240] loss: 1.530 [49, 300] loss: 1.523 [49, 360] loss: 1.535 Epoch: 49 -> Loss: 1.56974923611 Epoch: 49 -> Test Accuracy: 39.79 [50, 60] loss: 1.515 [50, 120] loss: 1.538 [50, 180] loss: 1.540 [50, 240] loss: 1.515 [50, 300] loss: 1.529 [50, 360] loss: 1.540 Epoch: 50 -> Loss: 1.53327488899 Epoch: 50 -> Test Accuracy: 38.71 [51, 60] loss: 1.525 [51, 120] loss: 1.540 [51, 180] loss: 1.522 [51, 240] loss: 1.542 [51, 300] loss: 1.527 [51, 360] loss: 1.529 Epoch: 51 -> Loss: 1.69156837463 Epoch: 51 -> Test Accuracy: 39.63 [52, 60] loss: 1.511 [52, 120] loss: 1.528 [52, 180] loss: 1.542 [52, 240] loss: 1.522 [52, 300] loss: 1.522 [52, 360] loss: 1.538 Epoch: 52 -> Loss: 1.38336443901 Epoch: 52 -> Test Accuracy: 39.22 [53, 60] loss: 1.530 [53, 120] loss: 1.517 [53, 180] loss: 1.533 [53, 240] loss: 1.524 [53, 300] loss: 1.526 [53, 360] loss: 1.509 Epoch: 53 -> Loss: 1.59752094746 Epoch: 53 -> Test Accuracy: 39.95 [54, 60] loss: 1.533 [54, 120] loss: 1.537 [54, 180] loss: 1.518 [54, 240] loss: 1.544 [54, 300] loss: 1.546 [54, 360] loss: 1.531 Epoch: 54 -> Loss: 1.40547215939 Epoch: 54 -> Test Accuracy: 38.85 [55, 60] loss: 1.525 [55, 120] loss: 1.537 [55, 180] loss: 1.525 [55, 240] loss: 1.515 [55, 300] loss: 1.520 [55, 360] loss: 1.539 Epoch: 55 -> Loss: 1.69401836395 Epoch: 55 -> Test Accuracy: 39.4 [56, 60] loss: 1.503 [56, 120] loss: 1.526 [56, 180] loss: 1.532 [56, 240] loss: 1.537 [56, 300] loss: 1.542 [56, 360] loss: 1.533 Epoch: 56 -> Loss: 1.69933640957 Epoch: 56 -> Test Accuracy: 37.49 [57, 60] loss: 1.534 [57, 120] loss: 1.539 [57, 180] loss: 1.534 [57, 240] loss: 1.530 [57, 300] loss: 1.529 [57, 360] loss: 1.532 Epoch: 57 -> Loss: 1.76949155331 Epoch: 57 -> Test Accuracy: 38.53 [58, 60] loss: 1.514 [58, 120] loss: 1.526 [58, 180] loss: 1.531 [58, 240] loss: 1.512 [58, 300] loss: 1.542 [58, 360] loss: 1.516 Epoch: 58 -> Loss: 1.4495652914 Epoch: 58 -> Test Accuracy: 39.72 [59, 60] loss: 1.521 [59, 120] loss: 1.541 [59, 180] loss: 1.526 [59, 240] loss: 1.526 [59, 300] loss: 1.519 [59, 360] loss: 1.524 Epoch: 59 -> Loss: 1.4615432024 Epoch: 59 -> Test Accuracy: 39.39 [60, 60] loss: 1.542 [60, 120] loss: 1.527 [60, 180] loss: 1.518 [60, 240] loss: 1.515 [60, 300] loss: 1.519 [60, 360] loss: 1.531 Epoch: 60 -> Loss: 1.45971381664 Epoch: 60 -> Test Accuracy: 39.49 [61, 60] loss: 1.521 [61, 120] loss: 1.517 [61, 180] loss: 1.515 [61, 240] loss: 1.509 [61, 300] loss: 1.531 [61, 360] loss: 1.527 Epoch: 61 -> Loss: 1.49451220036 Epoch: 61 -> Test Accuracy: 40.02 [62, 60] loss: 1.508 [62, 120] loss: 1.547 [62, 180] loss: 1.511 [62, 240] loss: 1.516 [62, 300] loss: 1.537 [62, 360] loss: 1.541 Epoch: 62 -> Loss: 1.34865796566 Epoch: 62 -> Test Accuracy: 38.87 [63, 60] loss: 1.545 [63, 120] loss: 1.524 [63, 180] loss: 1.520 [63, 240] loss: 1.530 [63, 300] loss: 1.533 [63, 360] loss: 1.512 Epoch: 63 -> Loss: 1.61831688881 Epoch: 63 -> Test Accuracy: 39.84 [64, 60] loss: 1.525 [64, 120] loss: 1.521 [64, 180] loss: 1.517 [64, 240] loss: 1.531 [64, 300] loss: 1.535 [64, 360] loss: 1.521 Epoch: 64 -> Loss: 1.5459883213 Epoch: 64 -> Test Accuracy: 39.62 [65, 60] loss: 1.520 [65, 120] loss: 1.531 [65, 180] loss: 1.526 [65, 240] loss: 1.501 [65, 300] loss: 1.517 [65, 360] loss: 1.519 Epoch: 65 -> Loss: 1.60374510288 Epoch: 65 -> Test Accuracy: 39.53 [66, 60] loss: 1.515 [66, 120] loss: 1.512 [66, 180] loss: 1.532 [66, 240] loss: 1.512 [66, 300] loss: 1.515 [66, 360] loss: 1.536 Epoch: 66 -> Loss: 1.51004242897 Epoch: 66 -> Test Accuracy: 39.36 [67, 60] loss: 1.514 [67, 120] loss: 1.515 [67, 180] loss: 1.528 [67, 240] loss: 1.521 [67, 300] loss: 1.518 [67, 360] loss: 1.527 Epoch: 67 -> Loss: 1.54676115513 Epoch: 67 -> Test Accuracy: 40.33 [68, 60] loss: 1.520 [68, 120] loss: 1.516 [68, 180] loss: 1.514 [68, 240] loss: 1.521 [68, 300] loss: 1.523 [68, 360] loss: 1.520 Epoch: 68 -> Loss: 1.45840704441 Epoch: 68 -> Test Accuracy: 40.35 [69, 60] loss: 1.508 [69, 120] loss: 1.525 [69, 180] loss: 1.530 [69, 240] loss: 1.525 [69, 300] loss: 1.518 [69, 360] loss: 1.496 Epoch: 69 -> Loss: 1.57058608532 Epoch: 69 -> Test Accuracy: 38.95 [70, 60] loss: 1.514 [70, 120] loss: 1.530 [70, 180] loss: 1.507 [70, 240] loss: 1.520 [70, 300] loss: 1.514 [70, 360] loss: 1.514 Epoch: 70 -> Loss: 1.31466925144 Epoch: 70 -> Test Accuracy: 38.88 [71, 60] loss: 1.489 [71, 120] loss: 1.456 [71, 180] loss: 1.453 [71, 240] loss: 1.453 [71, 300] loss: 1.453 [71, 360] loss: 1.449 Epoch: 71 -> Loss: 1.49236416817 Epoch: 71 -> Test Accuracy: 41.84 [72, 60] loss: 1.437 [72, 120] loss: 1.451 [72, 180] loss: 1.434 [72, 240] loss: 1.430 [72, 300] loss: 1.430 [72, 360] loss: 1.435 Epoch: 72 -> Loss: 1.4060229063 Epoch: 72 -> Test Accuracy: 42.13 [73, 60] loss: 1.443 [73, 120] loss: 1.424 [73, 180] loss: 1.436 [73, 240] loss: 1.435 [73, 300] loss: 1.444 [73, 360] loss: 1.436 Epoch: 73 -> Loss: 1.50561439991 Epoch: 73 -> Test Accuracy: 41.87 [74, 60] loss: 1.421 [74, 120] loss: 1.427 [74, 180] loss: 1.413 [74, 240] loss: 1.438 [74, 300] loss: 1.424 [74, 360] loss: 1.453 Epoch: 74 -> Loss: 1.41287982464 Epoch: 74 -> Test Accuracy: 42.1 [75, 60] loss: 1.424 [75, 120] loss: 1.431 [75, 180] loss: 1.435 [75, 240] loss: 1.432 [75, 300] loss: 1.438 [75, 360] loss: 1.422 Epoch: 75 -> Loss: 1.51846063137 Epoch: 75 -> Test Accuracy: 42.08 [76, 60] loss: 1.435 [76, 120] loss: 1.415 [76, 180] loss: 1.417 [76, 240] loss: 1.423 [76, 300] loss: 1.429 [76, 360] loss: 1.437 Epoch: 76 -> Loss: 1.38681399822 Epoch: 76 -> Test Accuracy: 41.99 [77, 60] loss: 1.432 [77, 120] loss: 1.415 [77, 180] loss: 1.419 [77, 240] loss: 1.412 [77, 300] loss: 1.433 [77, 360] loss: 1.422 Epoch: 77 -> Loss: 1.33687949181 Epoch: 77 -> Test Accuracy: 41.84 [78, 60] loss: 1.425 [78, 120] loss: 1.433 [78, 180] loss: 1.436 [78, 240] loss: 1.439 [78, 300] loss: 1.428 [78, 360] loss: 1.423 Epoch: 78 -> Loss: 1.44703269005 Epoch: 78 -> Test Accuracy: 42.31 [79, 60] loss: 1.423 [79, 120] loss: 1.419 [79, 180] loss: 1.409 [79, 240] loss: 1.442 [79, 300] loss: 1.407 [79, 360] loss: 1.421 Epoch: 79 -> Loss: 1.3942360878 Epoch: 79 -> Test Accuracy: 42.41 [80, 60] loss: 1.423 [80, 120] loss: 1.416 [80, 180] loss: 1.433 [80, 240] loss: 1.409 [80, 300] loss: 1.422 [80, 360] loss: 1.430 Epoch: 80 -> Loss: 1.42237174511 Epoch: 80 -> Test Accuracy: 41.98 [81, 60] loss: 1.421 [81, 120] loss: 1.427 [81, 180] loss: 1.438 [81, 240] loss: 1.431 [81, 300] loss: 1.414 [81, 360] loss: 1.402 Epoch: 81 -> Loss: 1.53787398338 Epoch: 81 -> Test Accuracy: 42.17 [82, 60] loss: 1.411 [82, 120] loss: 1.421 [82, 180] loss: 1.426 [82, 240] loss: 1.409 [82, 300] loss: 1.447 [82, 360] loss: 1.400 Epoch: 82 -> Loss: 1.53660416603 Epoch: 82 -> Test Accuracy: 42.45 [83, 60] loss: 1.420 [83, 120] loss: 1.415 [83, 180] loss: 1.432 [83, 240] loss: 1.410 [83, 300] loss: 1.421 [83, 360] loss: 1.418 Epoch: 83 -> Loss: 1.52950870991 Epoch: 83 -> Test Accuracy: 42.06 [84, 60] loss: 1.424 [84, 120] loss: 1.393 [84, 180] loss: 1.414 [84, 240] loss: 1.426 [84, 300] loss: 1.430 [84, 360] loss: 1.414 Epoch: 84 -> Loss: 1.31609892845 Epoch: 84 -> Test Accuracy: 41.9 [85, 60] loss: 1.404 [85, 120] loss: 1.409 [85, 180] loss: 1.432 [85, 240] loss: 1.419 [85, 300] loss: 1.414 [85, 360] loss: 1.417 Epoch: 85 -> Loss: 1.54289126396 Epoch: 85 -> Test Accuracy: 42.53 [86, 60] loss: 1.405 [86, 120] loss: 1.404 [86, 180] loss: 1.381 [86, 240] loss: 1.367 [86, 300] loss: 1.392 [86, 360] loss: 1.386 Epoch: 86 -> Loss: 1.40317416191 Epoch: 86 -> Test Accuracy: 43.37 [87, 60] loss: 1.387 [87, 120] loss: 1.392 [87, 180] loss: 1.383 [87, 240] loss: 1.388 [87, 300] loss: 1.392 [87, 360] loss: 1.384 Epoch: 87 -> Loss: 1.14942085743 Epoch: 87 -> Test Accuracy: 43.66 [88, 60] loss: 1.375 [88, 120] loss: 1.392 [88, 180] loss: 1.362 [88, 240] loss: 1.390 [88, 300] loss: 1.385 [88, 360] loss: 1.393 Epoch: 88 -> Loss: 1.40160965919 Epoch: 88 -> Test Accuracy: 43.63 [89, 60] loss: 1.380 [89, 120] loss: 1.368 [89, 180] loss: 1.383 [89, 240] loss: 1.389 [89, 300] loss: 1.370 [89, 360] loss: 1.382 Epoch: 89 -> Loss: 1.53526997566 Epoch: 89 -> Test Accuracy: 43.54 [90, 60] loss: 1.373 [90, 120] loss: 1.366 [90, 180] loss: 1.388 [90, 240] loss: 1.379 [90, 300] loss: 1.400 [90, 360] loss: 1.375 Epoch: 90 -> Loss: 1.29325318336 Epoch: 90 -> Test Accuracy: 43.5 [91, 60] loss: 1.378 [91, 120] loss: 1.367 [91, 180] loss: 1.393 [91, 240] loss: 1.391 [91, 300] loss: 1.387 [91, 360] loss: 1.379 Epoch: 91 -> Loss: 1.52791833878 Epoch: 91 -> Test Accuracy: 43.51 [92, 60] loss: 1.384 [92, 120] loss: 1.378 [92, 180] loss: 1.388 [92, 240] loss: 1.381 [92, 300] loss: 1.371 [92, 360] loss: 1.373 Epoch: 92 -> Loss: 1.41006708145 Epoch: 92 -> Test Accuracy: 43.55 [93, 60] loss: 1.382 [93, 120] loss: 1.377 [93, 180] loss: 1.377 [93, 240] loss: 1.377 [93, 300] loss: 1.393 [93, 360] loss: 1.392 Epoch: 93 -> Loss: 1.23832499981 Epoch: 93 -> Test Accuracy: 43.65 [94, 60] loss: 1.377 [94, 120] loss: 1.370 [94, 180] loss: 1.390 [94, 240] loss: 1.392 [94, 300] loss: 1.394 [94, 360] loss: 1.381 Epoch: 94 -> Loss: 1.4056199789 Epoch: 94 -> Test Accuracy: 43.52 [95, 60] loss: 1.385 [95, 120] loss: 1.379 [95, 180] loss: 1.367 [95, 240] loss: 1.388 [95, 300] loss: 1.377 [95, 360] loss: 1.363 Epoch: 95 -> Loss: 1.36497354507 Epoch: 95 -> Test Accuracy: 43.56 [96, 60] loss: 1.369 [96, 120] loss: 1.377 [96, 180] loss: 1.368 [96, 240] loss: 1.380 [96, 300] loss: 1.367 [96, 360] loss: 1.367 Epoch: 96 -> Loss: 1.27985405922 Epoch: 96 -> Test Accuracy: 43.78 [97, 60] loss: 1.359 [97, 120] loss: 1.386 [97, 180] loss: 1.391 [97, 240] loss: 1.372 [97, 300] loss: 1.385 [97, 360] loss: 1.356 Epoch: 97 -> Loss: 1.41945564747 Epoch: 97 -> Test Accuracy: 43.38 [98, 60] loss: 1.380 [98, 120] loss: 1.386 [98, 180] loss: 1.381 [98, 240] loss: 1.387 [98, 300] loss: 1.365 [98, 360] loss: 1.384 Epoch: 98 -> Loss: 1.42576658726 Epoch: 98 -> Test Accuracy: 43.45 [99, 60] loss: 1.368 [99, 120] loss: 1.370 [99, 180] loss: 1.392 [99, 240] loss: 1.380 [99, 300] loss: 1.370 [99, 360] loss: 1.386 Epoch: 99 -> Loss: 1.31827950478 Epoch: 99 -> Test Accuracy: 43.78 [100, 60] loss: 1.374 [100, 120] loss: 1.376 [100, 180] loss: 1.360 [100, 240] loss: 1.366 [100, 300] loss: 1.375 [100, 360] loss: 1.381 Epoch: 100 -> Loss: 1.44635605812 Epoch: 100 -> Test Accuracy: 43.9 Finished Training
In [10]:
# save variables
fm.save_variable([rot_block5_loss_log, rot_block5_test_accuracy_log,
block5_loss_log, block5_test_accuracy_log,
conv_block5_loss_log, conv_block5_test_accuracy_log], "5_block_net")
In [11]:
# rename files
fm.add_block_to_name(5, [100, 200])
Supervised NIN¶
Note: In the code of the paper a 3 convolutional block RotNet was used for the classification task.
In [9]:
# initialize networks
net_class = RN.RotNet(num_classes=10, num_conv_block=3, add_avg_pool=False)
In [10]:
# train 3 block RotNet on classification task
class_NIN_loss_log, _, class_NIN_test_accuracy_log, _, _ = tr.adaptive_learning([0.1, 0.02, 0.004, 0.0008],
[60, 120, 160, 200], 0.9, 5e-4, net_class, criterion, trainloader, None, testloader)
[1, 60] loss: 1.751 [1, 120] loss: 1.480 [1, 180] loss: 1.339 [1, 240] loss: 1.254 [1, 300] loss: 1.175 [1, 360] loss: 1.112 Epoch: 1 -> Loss: 0.825036227703 Epoch: 1 -> Test Accuracy: 60.62 [2, 60] loss: 1.038 [2, 120] loss: 0.998 [2, 180] loss: 0.963 [2, 240] loss: 0.930 [2, 300] loss: 0.911 [2, 360] loss: 0.883 Epoch: 2 -> Loss: 0.775784909725 Epoch: 2 -> Test Accuracy: 69.0 [3, 60] loss: 0.800 [3, 120] loss: 0.807 [3, 180] loss: 0.783 [3, 240] loss: 0.803 [3, 300] loss: 0.791 [3, 360] loss: 0.775 Epoch: 3 -> Loss: 0.804202079773 Epoch: 3 -> Test Accuracy: 72.79 [4, 60] loss: 0.708 [4, 120] loss: 0.718 [4, 180] loss: 0.737 [4, 240] loss: 0.714 [4, 300] loss: 0.705 [4, 360] loss: 0.708 Epoch: 4 -> Loss: 0.555963754654 Epoch: 4 -> Test Accuracy: 75.21 [5, 60] loss: 0.648 [5, 120] loss: 0.670 [5, 180] loss: 0.670 [5, 240] loss: 0.658 [5, 300] loss: 0.669 [5, 360] loss: 0.655 Epoch: 5 -> Loss: 0.522774040699 Epoch: 5 -> Test Accuracy: 75.59 [6, 60] loss: 0.632 [6, 120] loss: 0.638 [6, 180] loss: 0.635 [6, 240] loss: 0.631 [6, 300] loss: 0.636 [6, 360] loss: 0.617 Epoch: 6 -> Loss: 0.645114660263 Epoch: 6 -> Test Accuracy: 77.1 [7, 60] loss: 0.598 [7, 120] loss: 0.603 [7, 180] loss: 0.604 [7, 240] loss: 0.593 [7, 300] loss: 0.605 [7, 360] loss: 0.594 Epoch: 7 -> Loss: 0.616147696972 Epoch: 7 -> Test Accuracy: 77.37 [8, 60] loss: 0.588 [8, 120] loss: 0.577 [8, 180] loss: 0.567 [8, 240] loss: 0.556 [8, 300] loss: 0.587 [8, 360] loss: 0.568 Epoch: 8 -> Loss: 0.573764920235 Epoch: 8 -> Test Accuracy: 78.65 [9, 60] loss: 0.524 [9, 120] loss: 0.565 [9, 180] loss: 0.547 [9, 240] loss: 0.570 [9, 300] loss: 0.547 [9, 360] loss: 0.549 Epoch: 9 -> Loss: 0.599540233612 Epoch: 9 -> Test Accuracy: 79.55 [10, 60] loss: 0.528 [10, 120] loss: 0.548 [10, 180] loss: 0.523 [10, 240] loss: 0.549 [10, 300] loss: 0.543 [10, 360] loss: 0.568 Epoch: 10 -> Loss: 0.459079831839 Epoch: 10 -> Test Accuracy: 79.53 [11, 60] loss: 0.512 [11, 120] loss: 0.509 [11, 180] loss: 0.520 [11, 240] loss: 0.528 [11, 300] loss: 0.552 [11, 360] loss: 0.524 Epoch: 11 -> Loss: 0.526621758938 Epoch: 11 -> Test Accuracy: 79.39 [12, 60] loss: 0.503 [12, 120] loss: 0.503 [12, 180] loss: 0.491 [12, 240] loss: 0.546 [12, 300] loss: 0.506 [12, 360] loss: 0.537 Epoch: 12 -> Loss: 0.492984056473 Epoch: 12 -> Test Accuracy: 80.46 [13, 60] loss: 0.488 [13, 120] loss: 0.493 [13, 180] loss: 0.512 [13, 240] loss: 0.502 [13, 300] loss: 0.521 [13, 360] loss: 0.511 Epoch: 13 -> Loss: 0.88372194767 Epoch: 13 -> Test Accuracy: 80.23 [14, 60] loss: 0.489 [14, 120] loss: 0.492 [14, 180] loss: 0.496 [14, 240] loss: 0.493 [14, 300] loss: 0.504 [14, 360] loss: 0.482 Epoch: 14 -> Loss: 0.431423246861 Epoch: 14 -> Test Accuracy: 80.05 [15, 60] loss: 0.496 [15, 120] loss: 0.476 [15, 180] loss: 0.500 [15, 240] loss: 0.499 [15, 300] loss: 0.473 [15, 360] loss: 0.502 Epoch: 15 -> Loss: 0.477042138577 Epoch: 15 -> Test Accuracy: 80.66 [16, 60] loss: 0.460 [16, 120] loss: 0.495 [16, 180] loss: 0.456 [16, 240] loss: 0.480 [16, 300] loss: 0.463 [16, 360] loss: 0.492 Epoch: 16 -> Loss: 0.489493042231 Epoch: 16 -> Test Accuracy: 81.42 [17, 60] loss: 0.481 [17, 120] loss: 0.447 [17, 180] loss: 0.472 [17, 240] loss: 0.449 [17, 300] loss: 0.476 [17, 360] loss: 0.476 Epoch: 17 -> Loss: 0.366201579571 Epoch: 17 -> Test Accuracy: 81.96 [18, 60] loss: 0.459 [18, 120] loss: 0.470 [18, 180] loss: 0.490 [18, 240] loss: 0.457 [18, 300] loss: 0.471 [18, 360] loss: 0.464 Epoch: 18 -> Loss: 0.637535214424 Epoch: 18 -> Test Accuracy: 82.02 [19, 60] loss: 0.436 [19, 120] loss: 0.467 [19, 180] loss: 0.462 [19, 240] loss: 0.464 [19, 300] loss: 0.457 [19, 360] loss: 0.468 Epoch: 19 -> Loss: 0.519069314003 Epoch: 19 -> Test Accuracy: 82.46 [20, 60] loss: 0.462 [20, 120] loss: 0.441 [20, 180] loss: 0.461 [20, 240] loss: 0.458 [20, 300] loss: 0.445 [20, 360] loss: 0.462 Epoch: 20 -> Loss: 0.585631966591 Epoch: 20 -> Test Accuracy: 82.5 [21, 60] loss: 0.439 [21, 120] loss: 0.439 [21, 180] loss: 0.454 [21, 240] loss: 0.454 [21, 300] loss: 0.471 [21, 360] loss: 0.434 Epoch: 21 -> Loss: 0.443163454533 Epoch: 21 -> Test Accuracy: 81.27 [22, 60] loss: 0.407 [22, 120] loss: 0.436 [22, 180] loss: 0.473 [22, 240] loss: 0.449 [22, 300] loss: 0.457 [22, 360] loss: 0.440 Epoch: 22 -> Loss: 0.397458344698 Epoch: 22 -> Test Accuracy: 82.7 [23, 60] loss: 0.421 [23, 120] loss: 0.423 [23, 180] loss: 0.455 [23, 240] loss: 0.463 [23, 300] loss: 0.432 [23, 360] loss: 0.454 Epoch: 23 -> Loss: 0.573688149452 Epoch: 23 -> Test Accuracy: 81.06 [24, 60] loss: 0.425 [24, 120] loss: 0.408 [24, 180] loss: 0.464 [24, 240] loss: 0.438 [24, 300] loss: 0.448 [24, 360] loss: 0.435 Epoch: 24 -> Loss: 0.485114812851 Epoch: 24 -> Test Accuracy: 82.19 [25, 60] loss: 0.425 [25, 120] loss: 0.403 [25, 180] loss: 0.429 [25, 240] loss: 0.438 [25, 300] loss: 0.437 [25, 360] loss: 0.459 Epoch: 25 -> Loss: 0.451935589314 Epoch: 25 -> Test Accuracy: 82.87 [26, 60] loss: 0.432 [26, 120] loss: 0.423 [26, 180] loss: 0.420 [26, 240] loss: 0.435 [26, 300] loss: 0.446 [26, 360] loss: 0.417 Epoch: 26 -> Loss: 0.521824896336 Epoch: 26 -> Test Accuracy: 83.0 [27, 60] loss: 0.420 [27, 120] loss: 0.419 [27, 180] loss: 0.432 [27, 240] loss: 0.444 [27, 300] loss: 0.419 [27, 360] loss: 0.461 Epoch: 27 -> Loss: 0.448279857635 Epoch: 27 -> Test Accuracy: 82.64 [28, 60] loss: 0.402 [28, 120] loss: 0.403 [28, 180] loss: 0.417 [28, 240] loss: 0.444 [28, 300] loss: 0.426 [28, 360] loss: 0.435 Epoch: 28 -> Loss: 0.296328753233 Epoch: 28 -> Test Accuracy: 83.38 [29, 60] loss: 0.413 [29, 120] loss: 0.423 [29, 180] loss: 0.419 [29, 240] loss: 0.417 [29, 300] loss: 0.442 [29, 360] loss: 0.436 Epoch: 29 -> Loss: 0.414182603359 Epoch: 29 -> Test Accuracy: 82.96 [30, 60] loss: 0.396 [30, 120] loss: 0.418 [30, 180] loss: 0.409 [30, 240] loss: 0.423 [30, 300] loss: 0.435 [30, 360] loss: 0.423 Epoch: 30 -> Loss: 0.492857694626 Epoch: 30 -> Test Accuracy: 82.43 [31, 60] loss: 0.415 [31, 120] loss: 0.407 [31, 180] loss: 0.405 [31, 240] loss: 0.409 [31, 300] loss: 0.420 [31, 360] loss: 0.429 Epoch: 31 -> Loss: 0.408995449543 Epoch: 31 -> Test Accuracy: 82.86 [32, 60] loss: 0.411 [32, 120] loss: 0.404 [32, 180] loss: 0.407 [32, 240] loss: 0.423 [32, 300] loss: 0.411 [32, 360] loss: 0.425 Epoch: 32 -> Loss: 0.489699691534 Epoch: 32 -> Test Accuracy: 84.13 [33, 60] loss: 0.380 [33, 120] loss: 0.398 [33, 180] loss: 0.428 [33, 240] loss: 0.424 [33, 300] loss: 0.423 [33, 360] loss: 0.424 Epoch: 33 -> Loss: 0.340313047171 Epoch: 33 -> Test Accuracy: 82.46 [34, 60] loss: 0.399 [34, 120] loss: 0.399 [34, 180] loss: 0.419 [34, 240] loss: 0.405 [34, 300] loss: 0.422 [34, 360] loss: 0.435 Epoch: 34 -> Loss: 0.443888813257 Epoch: 34 -> Test Accuracy: 84.08 [35, 60] loss: 0.397 [35, 120] loss: 0.399 [35, 180] loss: 0.400 [35, 240] loss: 0.437 [35, 300] loss: 0.420 [35, 360] loss: 0.411 Epoch: 35 -> Loss: 0.550703644753 Epoch: 35 -> Test Accuracy: 84.29 [36, 60] loss: 0.378 [36, 120] loss: 0.415 [36, 180] loss: 0.401 [36, 240] loss: 0.413 [36, 300] loss: 0.418 [36, 360] loss: 0.405 Epoch: 36 -> Loss: 0.388389289379 Epoch: 36 -> Test Accuracy: 82.57 [37, 60] loss: 0.395 [37, 120] loss: 0.403 [37, 180] loss: 0.417 [37, 240] loss: 0.433 [37, 300] loss: 0.428 [37, 360] loss: 0.420 Epoch: 37 -> Loss: 0.444215625525 Epoch: 37 -> Test Accuracy: 83.28 [38, 60] loss: 0.387 [38, 120] loss: 0.392 [38, 180] loss: 0.388 [38, 240] loss: 0.398 [38, 300] loss: 0.419 [38, 360] loss: 0.420 Epoch: 38 -> Loss: 0.378732860088 Epoch: 38 -> Test Accuracy: 82.44 [39, 60] loss: 0.387 [39, 120] loss: 0.380 [39, 180] loss: 0.399 [39, 240] loss: 0.414 [39, 300] loss: 0.413 [39, 360] loss: 0.404 Epoch: 39 -> Loss: 0.441468238831 Epoch: 39 -> Test Accuracy: 83.92 [40, 60] loss: 0.391 [40, 120] loss: 0.384 [40, 180] loss: 0.414 [40, 240] loss: 0.404 [40, 300] loss: 0.423 [40, 360] loss: 0.417 Epoch: 40 -> Loss: 0.44526296854 Epoch: 40 -> Test Accuracy: 83.79 [41, 60] loss: 0.391 [41, 120] loss: 0.384 [41, 180] loss: 0.414 [41, 240] loss: 0.414 [41, 300] loss: 0.413 [41, 360] loss: 0.408 Epoch: 41 -> Loss: 0.379072278738 Epoch: 41 -> Test Accuracy: 83.75 [42, 60] loss: 0.379 [42, 120] loss: 0.404 [42, 180] loss: 0.400 [42, 240] loss: 0.424 [42, 300] loss: 0.414 [42, 360] loss: 0.398 Epoch: 42 -> Loss: 0.593099057674 Epoch: 42 -> Test Accuracy: 83.77 [43, 60] loss: 0.380 [43, 120] loss: 0.389 [43, 180] loss: 0.401 [43, 240] loss: 0.431 [43, 300] loss: 0.406 [43, 360] loss: 0.399 Epoch: 43 -> Loss: 0.20330825448 Epoch: 43 -> Test Accuracy: 83.59 [44, 60] loss: 0.359 [44, 120] loss: 0.397 [44, 180] loss: 0.400 [44, 240] loss: 0.403 [44, 300] loss: 0.405 [44, 360] loss: 0.407 Epoch: 44 -> Loss: 0.428490787745 Epoch: 44 -> Test Accuracy: 84.12 [45, 60] loss: 0.389 [45, 120] loss: 0.412 [45, 180] loss: 0.382 [45, 240] loss: 0.397 [45, 300] loss: 0.401 [45, 360] loss: 0.413 Epoch: 45 -> Loss: 0.486002355814 Epoch: 45 -> Test Accuracy: 83.85 [46, 60] loss: 0.398 [46, 120] loss: 0.382 [46, 180] loss: 0.416 [46, 240] loss: 0.384 [46, 300] loss: 0.406 [46, 360] loss: 0.422 Epoch: 46 -> Loss: 0.473516404629 Epoch: 46 -> Test Accuracy: 82.9 [47, 60] loss: 0.363 [47, 120] loss: 0.402 [47, 180] loss: 0.396 [47, 240] loss: 0.389 [47, 300] loss: 0.410 [47, 360] loss: 0.407 Epoch: 47 -> Loss: 0.30972841382 Epoch: 47 -> Test Accuracy: 83.93 [48, 60] loss: 0.361 [48, 120] loss: 0.376 [48, 180] loss: 0.410 [48, 240] loss: 0.422 [48, 300] loss: 0.400 [48, 360] loss: 0.403 Epoch: 48 -> Loss: 0.2999766469 Epoch: 48 -> Test Accuracy: 84.31 [49, 60] loss: 0.394 [49, 120] loss: 0.380 [49, 180] loss: 0.405 [49, 240] loss: 0.398 [49, 300] loss: 0.404 [49, 360] loss: 0.381 Epoch: 49 -> Loss: 0.502337992191 Epoch: 49 -> Test Accuracy: 82.65 [50, 60] loss: 0.372 [50, 120] loss: 0.378 [50, 180] loss: 0.397 [50, 240] loss: 0.400 [50, 300] loss: 0.384 [50, 360] loss: 0.408 Epoch: 50 -> Loss: 0.469843149185 Epoch: 50 -> Test Accuracy: 85.27 [51, 60] loss: 0.368 [51, 120] loss: 0.398 [51, 180] loss: 0.380 [51, 240] loss: 0.404 [51, 300] loss: 0.403 [51, 360] loss: 0.399 Epoch: 51 -> Loss: 0.491511195898 Epoch: 51 -> Test Accuracy: 84.14 [52, 60] loss: 0.367 [52, 120] loss: 0.372 [52, 180] loss: 0.385 [52, 240] loss: 0.417 [52, 300] loss: 0.391 [52, 360] loss: 0.391 Epoch: 52 -> Loss: 0.439156144857 Epoch: 52 -> Test Accuracy: 84.07 [53, 60] loss: 0.371 [53, 120] loss: 0.373 [53, 180] loss: 0.390 [53, 240] loss: 0.409 [53, 300] loss: 0.411 [53, 360] loss: 0.411 Epoch: 53 -> Loss: 0.396387606859 Epoch: 53 -> Test Accuracy: 83.2 [54, 60] loss: 0.372 [54, 120] loss: 0.362 [54, 180] loss: 0.393 [54, 240] loss: 0.407 [54, 300] loss: 0.382 [54, 360] loss: 0.420 Epoch: 54 -> Loss: 0.492816776037 Epoch: 54 -> Test Accuracy: 83.64 [55, 60] loss: 0.379 [55, 120] loss: 0.402 [55, 180] loss: 0.403 [55, 240] loss: 0.368 [55, 300] loss: 0.382 [55, 360] loss: 0.406 Epoch: 55 -> Loss: 0.495667219162 Epoch: 55 -> Test Accuracy: 84.79 [56, 60] loss: 0.371 [56, 120] loss: 0.376 [56, 180] loss: 0.375 [56, 240] loss: 0.386 [56, 300] loss: 0.389 [56, 360] loss: 0.409 Epoch: 56 -> Loss: 0.326176345348 Epoch: 56 -> Test Accuracy: 84.16 [57, 60] loss: 0.376 [57, 120] loss: 0.373 [57, 180] loss: 0.389 [57, 240] loss: 0.400 [57, 300] loss: 0.388 [57, 360] loss: 0.378 Epoch: 57 -> Loss: 0.519686937332 Epoch: 57 -> Test Accuracy: 83.23 [58, 60] loss: 0.384 [58, 120] loss: 0.358 [58, 180] loss: 0.385 [58, 240] loss: 0.382 [58, 300] loss: 0.399 [58, 360] loss: 0.410 Epoch: 58 -> Loss: 0.343048483133 Epoch: 58 -> Test Accuracy: 84.65 [59, 60] loss: 0.359 [59, 120] loss: 0.384 [59, 180] loss: 0.395 [59, 240] loss: 0.391 [59, 300] loss: 0.385 [59, 360] loss: 0.396 Epoch: 59 -> Loss: 0.461993128061 Epoch: 59 -> Test Accuracy: 84.73 [60, 60] loss: 0.362 [60, 120] loss: 0.391 [60, 180] loss: 0.386 [60, 240] loss: 0.398 [60, 300] loss: 0.393 [60, 360] loss: 0.380 Epoch: 60 -> Loss: 0.526801228523 Epoch: 60 -> Test Accuracy: 81.88 [61, 60] loss: 0.280 [61, 120] loss: 0.221 [61, 180] loss: 0.220 [61, 240] loss: 0.225 [61, 300] loss: 0.212 [61, 360] loss: 0.205 Epoch: 61 -> Loss: 0.132441371679 Epoch: 61 -> Test Accuracy: 89.27 [62, 60] loss: 0.168 [62, 120] loss: 0.172 [62, 180] loss: 0.173 [62, 240] loss: 0.174 [62, 300] loss: 0.188 [62, 360] loss: 0.181 Epoch: 62 -> Loss: 0.262564599514 Epoch: 62 -> Test Accuracy: 89.49 [63, 60] loss: 0.152 [63, 120] loss: 0.166 [63, 180] loss: 0.161 [63, 240] loss: 0.155 [63, 300] loss: 0.162 [63, 360] loss: 0.162 Epoch: 63 -> Loss: 0.164224550128 Epoch: 63 -> Test Accuracy: 89.29 [64, 60] loss: 0.137 [64, 120] loss: 0.147 [64, 180] loss: 0.144 [64, 240] loss: 0.157 [64, 300] loss: 0.156 [64, 360] loss: 0.149 Epoch: 64 -> Loss: 0.199184060097 Epoch: 64 -> Test Accuracy: 89.32 [65, 60] loss: 0.124 [65, 120] loss: 0.140 [65, 180] loss: 0.148 [65, 240] loss: 0.131 [65, 300] loss: 0.153 [65, 360] loss: 0.149 Epoch: 65 -> Loss: 0.127772569656 Epoch: 65 -> Test Accuracy: 89.05 [66, 60] loss: 0.118 [66, 120] loss: 0.131 [66, 180] loss: 0.138 [66, 240] loss: 0.133 [66, 300] loss: 0.139 [66, 360] loss: 0.148 Epoch: 66 -> Loss: 0.161133691669 Epoch: 66 -> Test Accuracy: 89.35 [67, 60] loss: 0.134 [67, 120] loss: 0.125 [67, 180] loss: 0.116 [67, 240] loss: 0.138 [67, 300] loss: 0.141 [67, 360] loss: 0.137 Epoch: 67 -> Loss: 0.153490871191 Epoch: 67 -> Test Accuracy: 89.13 [68, 60] loss: 0.126 [68, 120] loss: 0.121 [68, 180] loss: 0.128 [68, 240] loss: 0.133 [68, 300] loss: 0.129 [68, 360] loss: 0.137 Epoch: 68 -> Loss: 0.109140112996 Epoch: 68 -> Test Accuracy: 88.57 [69, 60] loss: 0.122 [69, 120] loss: 0.120 [69, 180] loss: 0.123 [69, 240] loss: 0.130 [69, 300] loss: 0.146 [69, 360] loss: 0.140 Epoch: 69 -> Loss: 0.161554858088 Epoch: 69 -> Test Accuracy: 89.36 [70, 60] loss: 0.120 [70, 120] loss: 0.112 [70, 180] loss: 0.129 [70, 240] loss: 0.130 [70, 300] loss: 0.132 [70, 360] loss: 0.138 Epoch: 70 -> Loss: 0.102998875082 Epoch: 70 -> Test Accuracy: 88.62 [71, 60] loss: 0.130 [71, 120] loss: 0.116 [71, 180] loss: 0.129 [71, 240] loss: 0.134 [71, 300] loss: 0.154 [71, 360] loss: 0.139 Epoch: 71 -> Loss: 0.152618929744 Epoch: 71 -> Test Accuracy: 88.61 [72, 60] loss: 0.118 [72, 120] loss: 0.119 [72, 180] loss: 0.134 [72, 240] loss: 0.139 [72, 300] loss: 0.149 [72, 360] loss: 0.160 Epoch: 72 -> Loss: 0.129081323743 Epoch: 72 -> Test Accuracy: 88.7 [73, 60] loss: 0.121 [73, 120] loss: 0.134 [73, 180] loss: 0.135 [73, 240] loss: 0.123 [73, 300] loss: 0.140 [73, 360] loss: 0.133 Epoch: 73 -> Loss: 0.137998253107 Epoch: 73 -> Test Accuracy: 88.36 [74, 60] loss: 0.118 [74, 120] loss: 0.131 [74, 180] loss: 0.118 [74, 240] loss: 0.142 [74, 300] loss: 0.147 [74, 360] loss: 0.139 Epoch: 74 -> Loss: 0.0809362605214 Epoch: 74 -> Test Accuracy: 88.13 [75, 60] loss: 0.124 [75, 120] loss: 0.141 [75, 180] loss: 0.150 [75, 240] loss: 0.143 [75, 300] loss: 0.158 [75, 360] loss: 0.156 Epoch: 75 -> Loss: 0.30938565731 Epoch: 75 -> Test Accuracy: 87.59 [76, 60] loss: 0.123 [76, 120] loss: 0.126 [76, 180] loss: 0.134 [76, 240] loss: 0.132 [76, 300] loss: 0.144 [76, 360] loss: 0.143 Epoch: 76 -> Loss: 0.158531919122 Epoch: 76 -> Test Accuracy: 88.79 [77, 60] loss: 0.121 [77, 120] loss: 0.130 [77, 180] loss: 0.139 [77, 240] loss: 0.139 [77, 300] loss: 0.147 [77, 360] loss: 0.150 Epoch: 77 -> Loss: 0.077156893909 Epoch: 77 -> Test Accuracy: 88.3 [78, 60] loss: 0.114 [78, 120] loss: 0.127 [78, 180] loss: 0.122 [78, 240] loss: 0.136 [78, 300] loss: 0.141 [78, 360] loss: 0.156 Epoch: 78 -> Loss: 0.336257785559 Epoch: 78 -> Test Accuracy: 87.83 [79, 60] loss: 0.147 [79, 120] loss: 0.124 [79, 180] loss: 0.135 [79, 240] loss: 0.141 [79, 300] loss: 0.143 [79, 360] loss: 0.153 Epoch: 79 -> Loss: 0.121548376977 Epoch: 79 -> Test Accuracy: 87.59 [80, 60] loss: 0.131 [80, 120] loss: 0.127 [80, 180] loss: 0.125 [80, 240] loss: 0.139 [80, 300] loss: 0.152 [80, 360] loss: 0.163 Epoch: 80 -> Loss: 0.0924355834723 Epoch: 80 -> Test Accuracy: 88.42 [81, 60] loss: 0.125 [81, 120] loss: 0.123 [81, 180] loss: 0.125 [81, 240] loss: 0.140 [81, 300] loss: 0.143 [81, 360] loss: 0.139 Epoch: 81 -> Loss: 0.144900158048 Epoch: 81 -> Test Accuracy: 89.16 [82, 60] loss: 0.118 [82, 120] loss: 0.128 [82, 180] loss: 0.138 [82, 240] loss: 0.148 [82, 300] loss: 0.148 [82, 360] loss: 0.133 Epoch: 82 -> Loss: 0.26424741745 Epoch: 82 -> Test Accuracy: 87.97 [83, 60] loss: 0.114 [83, 120] loss: 0.126 [83, 180] loss: 0.129 [83, 240] loss: 0.136 [83, 300] loss: 0.133 [83, 360] loss: 0.149 Epoch: 83 -> Loss: 0.19867387414 Epoch: 83 -> Test Accuracy: 88.52 [84, 60] loss: 0.114 [84, 120] loss: 0.121 [84, 180] loss: 0.132 [84, 240] loss: 0.132 [84, 300] loss: 0.142 [84, 360] loss: 0.158 Epoch: 84 -> Loss: 0.0979826822877 Epoch: 84 -> Test Accuracy: 88.32 [85, 60] loss: 0.120 [85, 120] loss: 0.130 [85, 180] loss: 0.140 [85, 240] loss: 0.146 [85, 300] loss: 0.142 [85, 360] loss: 0.143 Epoch: 85 -> Loss: 0.166171133518 Epoch: 85 -> Test Accuracy: 88.5 [86, 60] loss: 0.119 [86, 120] loss: 0.122 [86, 180] loss: 0.136 [86, 240] loss: 0.129 [86, 300] loss: 0.135 [86, 360] loss: 0.156 Epoch: 86 -> Loss: 0.151281192899 Epoch: 86 -> Test Accuracy: 88.61 [87, 60] loss: 0.120 [87, 120] loss: 0.109 [87, 180] loss: 0.132 [87, 240] loss: 0.128 [87, 300] loss: 0.143 [87, 360] loss: 0.142 Epoch: 87 -> Loss: 0.173461005092 Epoch: 87 -> Test Accuracy: 88.67 [88, 60] loss: 0.116 [88, 120] loss: 0.128 [88, 180] loss: 0.130 [88, 240] loss: 0.139 [88, 300] loss: 0.140 [88, 360] loss: 0.140 Epoch: 88 -> Loss: 0.225553706288 Epoch: 88 -> Test Accuracy: 88.07 [89, 60] loss: 0.113 [89, 120] loss: 0.125 [89, 180] loss: 0.127 [89, 240] loss: 0.136 [89, 300] loss: 0.139 [89, 360] loss: 0.137 Epoch: 89 -> Loss: 0.165683954954 Epoch: 89 -> Test Accuracy: 87.6 [90, 60] loss: 0.106 [90, 120] loss: 0.113 [90, 180] loss: 0.122 [90, 240] loss: 0.137 [90, 300] loss: 0.142 [90, 360] loss: 0.143 Epoch: 90 -> Loss: 0.0895455107093 Epoch: 90 -> Test Accuracy: 88.4 [91, 60] loss: 0.130 [91, 120] loss: 0.119 [91, 180] loss: 0.124 [91, 240] loss: 0.137 [91, 300] loss: 0.147 [91, 360] loss: 0.138 Epoch: 91 -> Loss: 0.221307352185 Epoch: 91 -> Test Accuracy: 87.88 [92, 60] loss: 0.111 [92, 120] loss: 0.126 [92, 180] loss: 0.133 [92, 240] loss: 0.129 [92, 300] loss: 0.136 [92, 360] loss: 0.138 Epoch: 92 -> Loss: 0.268699020147 Epoch: 92 -> Test Accuracy: 88.3 [93, 60] loss: 0.110 [93, 120] loss: 0.115 [93, 180] loss: 0.137 [93, 240] loss: 0.123 [93, 300] loss: 0.136 [93, 360] loss: 0.133 Epoch: 93 -> Loss: 0.0883329063654 Epoch: 93 -> Test Accuracy: 87.29 [94, 60] loss: 0.134 [94, 120] loss: 0.125 [94, 180] loss: 0.128 [94, 240] loss: 0.124 [94, 300] loss: 0.128 [94, 360] loss: 0.137 Epoch: 94 -> Loss: 0.0805646926165 Epoch: 94 -> Test Accuracy: 88.1 [95, 60] loss: 0.104 [95, 120] loss: 0.120 [95, 180] loss: 0.114 [95, 240] loss: 0.135 [95, 300] loss: 0.135 [95, 360] loss: 0.139 Epoch: 95 -> Loss: 0.18209400773 Epoch: 95 -> Test Accuracy: 88.31 [96, 60] loss: 0.116 [96, 120] loss: 0.110 [96, 180] loss: 0.122 [96, 240] loss: 0.127 [96, 300] loss: 0.136 [96, 360] loss: 0.130 Epoch: 96 -> Loss: 0.0532575622201 Epoch: 96 -> Test Accuracy: 88.14 [97, 60] loss: 0.124 [97, 120] loss: 0.130 [97, 180] loss: 0.111 [97, 240] loss: 0.112 [97, 300] loss: 0.125 [97, 360] loss: 0.142 Epoch: 97 -> Loss: 0.068568430841 Epoch: 97 -> Test Accuracy: 88.73 [98, 60] loss: 0.114 [98, 120] loss: 0.112 [98, 180] loss: 0.123 [98, 240] loss: 0.115 [98, 300] loss: 0.132 [98, 360] loss: 0.140 Epoch: 98 -> Loss: 0.0740065425634 Epoch: 98 -> Test Accuracy: 88.96 [99, 60] loss: 0.116 [99, 120] loss: 0.118 [99, 180] loss: 0.119 [99, 240] loss: 0.122 [99, 300] loss: 0.134 [99, 360] loss: 0.150 Epoch: 99 -> Loss: 0.247417330742 Epoch: 99 -> Test Accuracy: 87.67 [100, 60] loss: 0.129 [100, 120] loss: 0.113 [100, 180] loss: 0.112 [100, 240] loss: 0.136 [100, 300] loss: 0.129 [100, 360] loss: 0.130 Epoch: 100 -> Loss: 0.089723482728 Epoch: 100 -> Test Accuracy: 88.47 [101, 60] loss: 0.115 [101, 120] loss: 0.111 [101, 180] loss: 0.128 [101, 240] loss: 0.127 [101, 300] loss: 0.132 [101, 360] loss: 0.138 Epoch: 101 -> Loss: 0.0839704573154 Epoch: 101 -> Test Accuracy: 87.68 [102, 60] loss: 0.102 [102, 120] loss: 0.111 [102, 180] loss: 0.131 [102, 240] loss: 0.129 [102, 300] loss: 0.149 [102, 360] loss: 0.136 Epoch: 102 -> Loss: 0.243998855352 Epoch: 102 -> Test Accuracy: 88.61 [103, 60] loss: 0.108 [103, 120] loss: 0.103 [103, 180] loss: 0.110 [103, 240] loss: 0.125 [103, 300] loss: 0.127 [103, 360] loss: 0.141 Epoch: 103 -> Loss: 0.0751181691885 Epoch: 103 -> Test Accuracy: 88.23 [104, 60] loss: 0.104 [104, 120] loss: 0.108 [104, 180] loss: 0.118 [104, 240] loss: 0.121 [104, 300] loss: 0.128 [104, 360] loss: 0.125 Epoch: 104 -> Loss: 0.214813321829 Epoch: 104 -> Test Accuracy: 88.3 [105, 60] loss: 0.101 [105, 120] loss: 0.103 [105, 180] loss: 0.115 [105, 240] loss: 0.115 [105, 300] loss: 0.131 [105, 360] loss: 0.143 Epoch: 105 -> Loss: 0.0993434637785 Epoch: 105 -> Test Accuracy: 88.33 [106, 60] loss: 0.107 [106, 120] loss: 0.106 [106, 180] loss: 0.107 [106, 240] loss: 0.113 [106, 300] loss: 0.138 [106, 360] loss: 0.127 Epoch: 106 -> Loss: 0.0725574865937 Epoch: 106 -> Test Accuracy: 88.67 [107, 60] loss: 0.105 [107, 120] loss: 0.109 [107, 180] loss: 0.116 [107, 240] loss: 0.123 [107, 300] loss: 0.137 [107, 360] loss: 0.144 Epoch: 107 -> Loss: 0.144227355719 Epoch: 107 -> Test Accuracy: 88.15 [108, 60] loss: 0.112 [108, 120] loss: 0.117 [108, 180] loss: 0.124 [108, 240] loss: 0.120 [108, 300] loss: 0.118 [108, 360] loss: 0.136 Epoch: 108 -> Loss: 0.162149980664 Epoch: 108 -> Test Accuracy: 88.18 [109, 60] loss: 0.099 [109, 120] loss: 0.110 [109, 180] loss: 0.120 [109, 240] loss: 0.126 [109, 300] loss: 0.121 [109, 360] loss: 0.131 Epoch: 109 -> Loss: 0.138436213136 Epoch: 109 -> Test Accuracy: 88.29 [110, 60] loss: 0.106 [110, 120] loss: 0.102 [110, 180] loss: 0.112 [110, 240] loss: 0.108 [110, 300] loss: 0.118 [110, 360] loss: 0.121 Epoch: 110 -> Loss: 0.143770366907 Epoch: 110 -> Test Accuracy: 88.23 [111, 60] loss: 0.110 [111, 120] loss: 0.107 [111, 180] loss: 0.112 [111, 240] loss: 0.110 [111, 300] loss: 0.114 [111, 360] loss: 0.129 Epoch: 111 -> Loss: 0.0900486707687 Epoch: 111 -> Test Accuracy: 88.19 [112, 60] loss: 0.109 [112, 120] loss: 0.110 [112, 180] loss: 0.114 [112, 240] loss: 0.139 [112, 300] loss: 0.118 [112, 360] loss: 0.144 Epoch: 112 -> Loss: 0.154474571347 Epoch: 112 -> Test Accuracy: 88.19 [113, 60] loss: 0.117 [113, 120] loss: 0.109 [113, 180] loss: 0.108 [113, 240] loss: 0.112 [113, 300] loss: 0.122 [113, 360] loss: 0.144 Epoch: 113 -> Loss: 0.0912730693817 Epoch: 113 -> Test Accuracy: 88.45 [114, 60] loss: 0.110 [114, 120] loss: 0.099 [114, 180] loss: 0.115 [114, 240] loss: 0.118 [114, 300] loss: 0.119 [114, 360] loss: 0.124 Epoch: 114 -> Loss: 0.0848844274879 Epoch: 114 -> Test Accuracy: 88.53 [115, 60] loss: 0.117 [115, 120] loss: 0.120 [115, 180] loss: 0.108 [115, 240] loss: 0.121 [115, 300] loss: 0.129 [115, 360] loss: 0.141 Epoch: 115 -> Loss: 0.103536307812 Epoch: 115 -> Test Accuracy: 88.43 [116, 60] loss: 0.105 [116, 120] loss: 0.105 [116, 180] loss: 0.112 [116, 240] loss: 0.122 [116, 300] loss: 0.122 [116, 360] loss: 0.126 Epoch: 116 -> Loss: 0.16309531033 Epoch: 116 -> Test Accuracy: 87.61 [117, 60] loss: 0.103 [117, 120] loss: 0.114 [117, 180] loss: 0.112 [117, 240] loss: 0.111 [117, 300] loss: 0.123 [117, 360] loss: 0.138 Epoch: 117 -> Loss: 0.0620582923293 Epoch: 117 -> Test Accuracy: 88.72 [118, 60] loss: 0.103 [118, 120] loss: 0.113 [118, 180] loss: 0.116 [118, 240] loss: 0.108 [118, 300] loss: 0.116 [118, 360] loss: 0.115 Epoch: 118 -> Loss: 0.087500795722 Epoch: 118 -> Test Accuracy: 87.27 [119, 60] loss: 0.106 [119, 120] loss: 0.114 [119, 180] loss: 0.126 [119, 240] loss: 0.104 [119, 300] loss: 0.103 [119, 360] loss: 0.122 Epoch: 119 -> Loss: 0.0654944702983 Epoch: 119 -> Test Accuracy: 88.68 [120, 60] loss: 0.111 [120, 120] loss: 0.101 [120, 180] loss: 0.114 [120, 240] loss: 0.118 [120, 300] loss: 0.132 [120, 360] loss: 0.132 Epoch: 120 -> Loss: 0.14968085289 Epoch: 120 -> Test Accuracy: 88.67 [121, 60] loss: 0.075 [121, 120] loss: 0.054 [121, 180] loss: 0.051 [121, 240] loss: 0.042 [121, 300] loss: 0.044 [121, 360] loss: 0.042 Epoch: 121 -> Loss: 0.0615567862988 Epoch: 121 -> Test Accuracy: 91.2 [122, 60] loss: 0.031 [122, 120] loss: 0.034 [122, 180] loss: 0.030 [122, 240] loss: 0.029 [122, 300] loss: 0.031 [122, 360] loss: 0.032 Epoch: 122 -> Loss: 0.0204253550619 Epoch: 122 -> Test Accuracy: 91.56 [123, 60] loss: 0.025 [123, 120] loss: 0.026 [123, 180] loss: 0.026 [123, 240] loss: 0.026 [123, 300] loss: 0.025 [123, 360] loss: 0.026 Epoch: 123 -> Loss: 0.0279893390834 Epoch: 123 -> Test Accuracy: 91.57 [124, 60] loss: 0.022 [124, 120] loss: 0.022 [124, 180] loss: 0.021 [124, 240] loss: 0.020 [124, 300] loss: 0.020 [124, 360] loss: 0.022 Epoch: 124 -> Loss: 0.00913160480559 Epoch: 124 -> Test Accuracy: 91.3 [125, 60] loss: 0.019 [125, 120] loss: 0.019 [125, 180] loss: 0.019 [125, 240] loss: 0.020 [125, 300] loss: 0.020 [125, 360] loss: 0.021 Epoch: 125 -> Loss: 0.0156837590039 Epoch: 125 -> Test Accuracy: 91.25 [126, 60] loss: 0.020 [126, 120] loss: 0.019 [126, 180] loss: 0.019 [126, 240] loss: 0.018 [126, 300] loss: 0.015 [126, 360] loss: 0.020 Epoch: 126 -> Loss: 0.00777263054624 Epoch: 126 -> Test Accuracy: 91.32 [127, 60] loss: 0.016 [127, 120] loss: 0.017 [127, 180] loss: 0.016 [127, 240] loss: 0.016 [127, 300] loss: 0.018 [127, 360] loss: 0.017 Epoch: 127 -> Loss: 0.0105040315539 Epoch: 127 -> Test Accuracy: 91.21 [128, 60] loss: 0.016 [128, 120] loss: 0.016 [128, 180] loss: 0.016 [128, 240] loss: 0.015 [128, 300] loss: 0.015 [128, 360] loss: 0.015 Epoch: 128 -> Loss: 0.0326036699116 Epoch: 128 -> Test Accuracy: 91.28 [129, 60] loss: 0.016 [129, 120] loss: 0.015 [129, 180] loss: 0.015 [129, 240] loss: 0.015 [129, 300] loss: 0.016 [129, 360] loss: 0.016 Epoch: 129 -> Loss: 0.0272331349552 Epoch: 129 -> Test Accuracy: 91.44 [130, 60] loss: 0.014 [130, 120] loss: 0.014 [130, 180] loss: 0.014 [130, 240] loss: 0.016 [130, 300] loss: 0.015 [130, 360] loss: 0.015 Epoch: 130 -> Loss: 0.0103903533891 Epoch: 130 -> Test Accuracy: 91.29 [131, 60] loss: 0.013 [131, 120] loss: 0.014 [131, 180] loss: 0.014 [131, 240] loss: 0.014 [131, 300] loss: 0.013 [131, 360] loss: 0.015 Epoch: 131 -> Loss: 0.0119562689215 Epoch: 131 -> Test Accuracy: 91.25 [132, 60] loss: 0.013 [132, 120] loss: 0.015 [132, 180] loss: 0.013 [132, 240] loss: 0.014 [132, 300] loss: 0.013 [132, 360] loss: 0.014 Epoch: 132 -> Loss: 0.0209854543209 Epoch: 132 -> Test Accuracy: 91.25 [133, 60] loss: 0.014 [133, 120] loss: 0.013 [133, 180] loss: 0.012 [133, 240] loss: 0.013 [133, 300] loss: 0.012 [133, 360] loss: 0.012 Epoch: 133 -> Loss: 0.0246900375932 Epoch: 133 -> Test Accuracy: 91.32 [134, 60] loss: 0.012 [134, 120] loss: 0.011 [134, 180] loss: 0.013 [134, 240] loss: 0.012 [134, 300] loss: 0.012 [134, 360] loss: 0.014 Epoch: 134 -> Loss: 0.014113759622 Epoch: 134 -> Test Accuracy: 91.12 [135, 60] loss: 0.011 [135, 120] loss: 0.010 [135, 180] loss: 0.012 [135, 240] loss: 0.012 [135, 300] loss: 0.011 [135, 360] loss: 0.011 Epoch: 135 -> Loss: 0.0275889132172 Epoch: 135 -> Test Accuracy: 91.17 [136, 60] loss: 0.012 [136, 120] loss: 0.012 [136, 180] loss: 0.010 [136, 240] loss: 0.011 [136, 300] loss: 0.011 [136, 360] loss: 0.011 Epoch: 136 -> Loss: 0.0197756737471 Epoch: 136 -> Test Accuracy: 91.07 [137, 60] loss: 0.011 [137, 120] loss: 0.011 [137, 180] loss: 0.011 [137, 240] loss: 0.011 [137, 300] loss: 0.010 [137, 360] loss: 0.011 Epoch: 137 -> Loss: 0.0155540648848 Epoch: 137 -> Test Accuracy: 91.52 [138, 60] loss: 0.011 [138, 120] loss: 0.011 [138, 180] loss: 0.011 [138, 240] loss: 0.010 [138, 300] loss: 0.011 [138, 360] loss: 0.010 Epoch: 138 -> Loss: 0.00844640098512 Epoch: 138 -> Test Accuracy: 91.52 [139, 60] loss: 0.010 [139, 120] loss: 0.010 [139, 180] loss: 0.011 [139, 240] loss: 0.011 [139, 300] loss: 0.011 [139, 360] loss: 0.010 Epoch: 139 -> Loss: 0.0280493143946 Epoch: 139 -> Test Accuracy: 91.36 [140, 60] loss: 0.011 [140, 120] loss: 0.010 [140, 180] loss: 0.012 [140, 240] loss: 0.011 [140, 300] loss: 0.012 [140, 360] loss: 0.011 Epoch: 140 -> Loss: 0.00766073446721 Epoch: 140 -> Test Accuracy: 91.4 [141, 60] loss: 0.010 [141, 120] loss: 0.010 [141, 180] loss: 0.010 [141, 240] loss: 0.011 [141, 300] loss: 0.011 [141, 360] loss: 0.010 Epoch: 141 -> Loss: 0.0104625821114 Epoch: 141 -> Test Accuracy: 91.31 [142, 60] loss: 0.011 [142, 120] loss: 0.009 [142, 180] loss: 0.010 [142, 240] loss: 0.012 [142, 300] loss: 0.009 [142, 360] loss: 0.011 Epoch: 142 -> Loss: 0.00761454086751 Epoch: 142 -> Test Accuracy: 91.44 [143, 60] loss: 0.010 [143, 120] loss: 0.009 [143, 180] loss: 0.010 [143, 240] loss: 0.010 [143, 300] loss: 0.011 [143, 360] loss: 0.010 Epoch: 143 -> Loss: 0.0193831622601 Epoch: 143 -> Test Accuracy: 91.42 [144, 60] loss: 0.011 [144, 120] loss: 0.010 [144, 180] loss: 0.009 [144, 240] loss: 0.009 [144, 300] loss: 0.010 [144, 360] loss: 0.009 Epoch: 144 -> Loss: 0.00670862803236 Epoch: 144 -> Test Accuracy: 91.32 [145, 60] loss: 0.010 [145, 120] loss: 0.010 [145, 180] loss: 0.009 [145, 240] loss: 0.010 [145, 300] loss: 0.011 [145, 360] loss: 0.011 Epoch: 145 -> Loss: 0.0103610632941 Epoch: 145 -> Test Accuracy: 91.4 [146, 60] loss: 0.010 [146, 120] loss: 0.009 [146, 180] loss: 0.009 [146, 240] loss: 0.010 [146, 300] loss: 0.009 [146, 360] loss: 0.009 Epoch: 146 -> Loss: 0.00978915672749 Epoch: 146 -> Test Accuracy: 91.5 [147, 60] loss: 0.009 [147, 120] loss: 0.010 [147, 180] loss: 0.009 [147, 240] loss: 0.009 [147, 300] loss: 0.010 [147, 360] loss: 0.009 Epoch: 147 -> Loss: 0.0133479358628 Epoch: 147 -> Test Accuracy: 91.39 [148, 60] loss: 0.009 [148, 120] loss: 0.009 [148, 180] loss: 0.009 [148, 240] loss: 0.009 [148, 300] loss: 0.008 [148, 360] loss: 0.009 Epoch: 148 -> Loss: 0.0146847125143 Epoch: 148 -> Test Accuracy: 91.49 [149, 60] loss: 0.008 [149, 120] loss: 0.009 [149, 180] loss: 0.010 [149, 240] loss: 0.009 [149, 300] loss: 0.009 [149, 360] loss: 0.009 Epoch: 149 -> Loss: 0.0182967539877 Epoch: 149 -> Test Accuracy: 91.29 [150, 60] loss: 0.009 [150, 120] loss: 0.009 [150, 180] loss: 0.008 [150, 240] loss: 0.008 [150, 300] loss: 0.009 [150, 360] loss: 0.010 Epoch: 150 -> Loss: 0.0112588824704 Epoch: 150 -> Test Accuracy: 91.47 [151, 60] loss: 0.010 [151, 120] loss: 0.009 [151, 180] loss: 0.009 [151, 240] loss: 0.009 [151, 300] loss: 0.009 [151, 360] loss: 0.009 Epoch: 151 -> Loss: 0.00795558094978 Epoch: 151 -> Test Accuracy: 91.39 [152, 60] loss: 0.009 [152, 120] loss: 0.009 [152, 180] loss: 0.008 [152, 240] loss: 0.008 [152, 300] loss: 0.009 [152, 360] loss: 0.009 Epoch: 152 -> Loss: 0.00826269388199 Epoch: 152 -> Test Accuracy: 91.09 [153, 60] loss: 0.009 [153, 120] loss: 0.010 [153, 180] loss: 0.009 [153, 240] loss: 0.008 [153, 300] loss: 0.009 [153, 360] loss: 0.009 Epoch: 153 -> Loss: 0.0122916577384 Epoch: 153 -> Test Accuracy: 91.09 [154, 60] loss: 0.009 [154, 120] loss: 0.008 [154, 180] loss: 0.008 [154, 240] loss: 0.008 [154, 300] loss: 0.009 [154, 360] loss: 0.009 Epoch: 154 -> Loss: 0.0158297717571 Epoch: 154 -> Test Accuracy: 91.09 [155, 60] loss: 0.008 [155, 120] loss: 0.009 [155, 180] loss: 0.010 [155, 240] loss: 0.009 [155, 300] loss: 0.009 [155, 360] loss: 0.009 Epoch: 155 -> Loss: 0.0062848450616 Epoch: 155 -> Test Accuracy: 91.14 [156, 60] loss: 0.009 [156, 120] loss: 0.008 [156, 180] loss: 0.009 [156, 240] loss: 0.009 [156, 300] loss: 0.009 [156, 360] loss: 0.009 Epoch: 156 -> Loss: 0.00874519906938 Epoch: 156 -> Test Accuracy: 91.34 [157, 60] loss: 0.009 [157, 120] loss: 0.008 [157, 180] loss: 0.009 [157, 240] loss: 0.008 [157, 300] loss: 0.009 [157, 360] loss: 0.008 Epoch: 157 -> Loss: 0.0130544928834 Epoch: 157 -> Test Accuracy: 91.1 [158, 60] loss: 0.008 [158, 120] loss: 0.009 [158, 180] loss: 0.008 [158, 240] loss: 0.009 [158, 300] loss: 0.008 [158, 360] loss: 0.009 Epoch: 158 -> Loss: 0.0168238095939 Epoch: 158 -> Test Accuracy: 91.31 [159, 60] loss: 0.009 [159, 120] loss: 0.008 [159, 180] loss: 0.008 [159, 240] loss: 0.008 [159, 300] loss: 0.008 [159, 360] loss: 0.010 Epoch: 159 -> Loss: 0.00392565131187 Epoch: 159 -> Test Accuracy: 91.2 [160, 60] loss: 0.008 [160, 120] loss: 0.008 [160, 180] loss: 0.008 [160, 240] loss: 0.008 [160, 300] loss: 0.009 [160, 360] loss: 0.008 Epoch: 160 -> Loss: 0.00465139141306 Epoch: 160 -> Test Accuracy: 91.41 [161, 60] loss: 0.007 [161, 120] loss: 0.007 [161, 180] loss: 0.007 [161, 240] loss: 0.008 [161, 300] loss: 0.008 [161, 360] loss: 0.007 Epoch: 161 -> Loss: 0.00661711674184 Epoch: 161 -> Test Accuracy: 91.22 [162, 60] loss: 0.008 [162, 120] loss: 0.007 [162, 180] loss: 0.007 [162, 240] loss: 0.007 [162, 300] loss: 0.007 [162, 360] loss: 0.007 Epoch: 162 -> Loss: 0.00278571853414 Epoch: 162 -> Test Accuracy: 91.34 [163, 60] loss: 0.007 [163, 120] loss: 0.007 [163, 180] loss: 0.007 [163, 240] loss: 0.008 [163, 300] loss: 0.007 [163, 360] loss: 0.007 Epoch: 163 -> Loss: 0.012463092804 Epoch: 163 -> Test Accuracy: 91.39 [164, 60] loss: 0.006 [164, 120] loss: 0.007 [164, 180] loss: 0.007 [164, 240] loss: 0.006 [164, 300] loss: 0.007 [164, 360] loss: 0.007 Epoch: 164 -> Loss: 0.03215944767 Epoch: 164 -> Test Accuracy: 91.28 [165, 60] loss: 0.007 [165, 120] loss: 0.006 [165, 180] loss: 0.007 [165, 240] loss: 0.006 [165, 300] loss: 0.007 [165, 360] loss: 0.007 Epoch: 165 -> Loss: 0.0114584919065 Epoch: 165 -> Test Accuracy: 91.31 [166, 60] loss: 0.007 [166, 120] loss: 0.006 [166, 180] loss: 0.006 [166, 240] loss: 0.007 [166, 300] loss: 0.007 [166, 360] loss: 0.007 Epoch: 166 -> Loss: 0.00599935650826 Epoch: 166 -> Test Accuracy: 91.39 [167, 60] loss: 0.007 [167, 120] loss: 0.006 [167, 180] loss: 0.007 [167, 240] loss: 0.006 [167, 300] loss: 0.006 [167, 360] loss: 0.007 Epoch: 167 -> Loss: 0.00732953567058 Epoch: 167 -> Test Accuracy: 91.35 [168, 60] loss: 0.006 [168, 120] loss: 0.007 [168, 180] loss: 0.006 [168, 240] loss: 0.006 [168, 300] loss: 0.006 [168, 360] loss: 0.007 Epoch: 168 -> Loss: 0.0101614054292 Epoch: 168 -> Test Accuracy: 91.38 [169, 60] loss: 0.007 [169, 120] loss: 0.006 [169, 180] loss: 0.006 [169, 240] loss: 0.006 [169, 300] loss: 0.007 [169, 360] loss: 0.007 Epoch: 169 -> Loss: 0.0114369038492 Epoch: 169 -> Test Accuracy: 91.44 [170, 60] loss: 0.007 [170, 120] loss: 0.007 [170, 180] loss: 0.007 [170, 240] loss: 0.007 [170, 300] loss: 0.007 [170, 360] loss: 0.007 Epoch: 170 -> Loss: 0.0217247419059 Epoch: 170 -> Test Accuracy: 91.27 [171, 60] loss: 0.007 [171, 120] loss: 0.006 [171, 180] loss: 0.006 [171, 240] loss: 0.007 [171, 300] loss: 0.007 [171, 360] loss: 0.007 Epoch: 171 -> Loss: 0.00443744659424 Epoch: 171 -> Test Accuracy: 91.35 [172, 60] loss: 0.006 [172, 120] loss: 0.007 [172, 180] loss: 0.006 [172, 240] loss: 0.007 [172, 300] loss: 0.006 [172, 360] loss: 0.006 Epoch: 172 -> Loss: 0.0275138020515 Epoch: 172 -> Test Accuracy: 91.45 [173, 60] loss: 0.006 [173, 120] loss: 0.007 [173, 180] loss: 0.006 [173, 240] loss: 0.006 [173, 300] loss: 0.006 [173, 360] loss: 0.006 Epoch: 173 -> Loss: 0.00900883041322 Epoch: 173 -> Test Accuracy: 91.31 [174, 60] loss: 0.006 [174, 120] loss: 0.007 [174, 180] loss: 0.006 [174, 240] loss: 0.006 [174, 300] loss: 0.007 [174, 360] loss: 0.007 Epoch: 174 -> Loss: 0.00365487346426 Epoch: 174 -> Test Accuracy: 91.32 [175, 60] loss: 0.006 [175, 120] loss: 0.006 [175, 180] loss: 0.006 [175, 240] loss: 0.007 [175, 300] loss: 0.006 [175, 360] loss: 0.006 Epoch: 175 -> Loss: 0.00441812258214 Epoch: 175 -> Test Accuracy: 91.28 [176, 60] loss: 0.006 [176, 120] loss: 0.007 [176, 180] loss: 0.006 [176, 240] loss: 0.006 [176, 300] loss: 0.006 [176, 360] loss: 0.006 Epoch: 176 -> Loss: 0.00915683526546 Epoch: 176 -> Test Accuracy: 91.42 [177, 60] loss: 0.007 [177, 120] loss: 0.007 [177, 180] loss: 0.006 [177, 240] loss: 0.006 [177, 300] loss: 0.007 [177, 360] loss: 0.007 Epoch: 177 -> Loss: 0.00952087063342 Epoch: 177 -> Test Accuracy: 91.3 [178, 60] loss: 0.007 [178, 120] loss: 0.006 [178, 180] loss: 0.006 [178, 240] loss: 0.007 [178, 300] loss: 0.007 [178, 360] loss: 0.007 Epoch: 178 -> Loss: 0.00619597453624 Epoch: 178 -> Test Accuracy: 91.36 [179, 60] loss: 0.006 [179, 120] loss: 0.006 [179, 180] loss: 0.006 [179, 240] loss: 0.007 [179, 300] loss: 0.007 [179, 360] loss: 0.007 Epoch: 179 -> Loss: 0.0256325509399 Epoch: 179 -> Test Accuracy: 91.39 [180, 60] loss: 0.007 [180, 120] loss: 0.006 [180, 180] loss: 0.007 [180, 240] loss: 0.007 [180, 300] loss: 0.006 [180, 360] loss: 0.007 Epoch: 180 -> Loss: 0.015832144767 Epoch: 180 -> Test Accuracy: 91.37 [181, 60] loss: 0.006 [181, 120] loss: 0.007 [181, 180] loss: 0.006 [181, 240] loss: 0.007 [181, 300] loss: 0.006 [181, 360] loss: 0.006 Epoch: 181 -> Loss: 0.00802031159401 Epoch: 181 -> Test Accuracy: 91.45 [182, 60] loss: 0.006 [182, 120] loss: 0.006 [182, 180] loss: 0.007 [182, 240] loss: 0.007 [182, 300] loss: 0.007 [182, 360] loss: 0.006 Epoch: 182 -> Loss: 0.00346255302429 Epoch: 182 -> Test Accuracy: 91.33 [183, 60] loss: 0.007 [183, 120] loss: 0.006 [183, 180] loss: 0.007 [183, 240] loss: 0.006 [183, 300] loss: 0.007 [183, 360] loss: 0.007 Epoch: 183 -> Loss: 0.00269183516502 Epoch: 183 -> Test Accuracy: 91.29 [184, 60] loss: 0.006 [184, 120] loss: 0.007 [184, 180] loss: 0.007 [184, 240] loss: 0.006 [184, 300] loss: 0.006 [184, 360] loss: 0.007 Epoch: 184 -> Loss: 0.00428304076195 Epoch: 184 -> Test Accuracy: 91.36 [185, 60] loss: 0.006 [185, 120] loss: 0.006 [185, 180] loss: 0.006 [185, 240] loss: 0.005 [185, 300] loss: 0.007 [185, 360] loss: 0.006 Epoch: 185 -> Loss: 0.0199624300003 Epoch: 185 -> Test Accuracy: 91.39 [186, 60] loss: 0.006 [186, 120] loss: 0.007 [186, 180] loss: 0.006 [186, 240] loss: 0.007 [186, 300] loss: 0.006 [186, 360] loss: 0.007 Epoch: 186 -> Loss: 0.00353973498568 Epoch: 186 -> Test Accuracy: 91.43 [187, 60] loss: 0.006 [187, 120] loss: 0.007 [187, 180] loss: 0.007 [187, 240] loss: 0.006 [187, 300] loss: 0.006 [187, 360] loss: 0.006 Epoch: 187 -> Loss: 0.00540779810399 Epoch: 187 -> Test Accuracy: 91.32 [188, 60] loss: 0.006 [188, 120] loss: 0.006 [188, 180] loss: 0.006 [188, 240] loss: 0.007 [188, 300] loss: 0.006 [188, 360] loss: 0.006 Epoch: 188 -> Loss: 0.0149178653955 Epoch: 188 -> Test Accuracy: 91.22 [189, 60] loss: 0.007 [189, 120] loss: 0.006 [189, 180] loss: 0.006 [189, 240] loss: 0.006 [189, 300] loss: 0.007 [189, 360] loss: 0.006 Epoch: 189 -> Loss: 0.00651356577873 Epoch: 189 -> Test Accuracy: 91.27 [190, 60] loss: 0.006 [190, 120] loss: 0.007 [190, 180] loss: 0.006 [190, 240] loss: 0.006 [190, 300] loss: 0.006 [190, 360] loss: 0.007 Epoch: 190 -> Loss: 0.00353185529821 Epoch: 190 -> Test Accuracy: 91.45 [191, 60] loss: 0.006 [191, 120] loss: 0.006 [191, 180] loss: 0.006 [191, 240] loss: 0.006 [191, 300] loss: 0.006 [191, 360] loss: 0.006 Epoch: 191 -> Loss: 0.00702169537544 Epoch: 191 -> Test Accuracy: 91.41 [192, 60] loss: 0.006 [192, 120] loss: 0.006 [192, 180] loss: 0.006 [192, 240] loss: 0.007 [192, 300] loss: 0.006 [192, 360] loss: 0.006 Epoch: 192 -> Loss: 0.00991275347769 Epoch: 192 -> Test Accuracy: 91.21 [193, 60] loss: 0.006 [193, 120] loss: 0.006 [193, 180] loss: 0.007 [193, 240] loss: 0.006 [193, 300] loss: 0.006 [193, 360] loss: 0.006 Epoch: 193 -> Loss: 0.00395756959915 Epoch: 193 -> Test Accuracy: 91.26 [194, 60] loss: 0.006 [194, 120] loss: 0.006 [194, 180] loss: 0.005 [194, 240] loss: 0.007 [194, 300] loss: 0.006 [194, 360] loss: 0.006 Epoch: 194 -> Loss: 0.00577930826694 Epoch: 194 -> Test Accuracy: 91.35 [195, 60] loss: 0.006 [195, 120] loss: 0.006 [195, 180] loss: 0.006 [195, 240] loss: 0.007 [195, 300] loss: 0.006 [195, 360] loss: 0.007 Epoch: 195 -> Loss: 0.00489818444476 Epoch: 195 -> Test Accuracy: 91.38 [196, 60] loss: 0.006 [196, 120] loss: 0.006 [196, 180] loss: 0.006 [196, 240] loss: 0.006 [196, 300] loss: 0.006 [196, 360] loss: 0.006 Epoch: 196 -> Loss: 0.0073150456883 Epoch: 196 -> Test Accuracy: 91.33 [197, 60] loss: 0.006 [197, 120] loss: 0.007 [197, 180] loss: 0.006 [197, 240] loss: 0.006 [197, 300] loss: 0.006 [197, 360] loss: 0.006 Epoch: 197 -> Loss: 0.017292862758 Epoch: 197 -> Test Accuracy: 91.25 [198, 60] loss: 0.006 [198, 120] loss: 0.006 [198, 180] loss: 0.006 [198, 240] loss: 0.006 [198, 300] loss: 0.006 [198, 360] loss: 0.007 Epoch: 198 -> Loss: 0.00433322181925 Epoch: 198 -> Test Accuracy: 91.31 [199, 60] loss: 0.006 [199, 120] loss: 0.006 [199, 180] loss: 0.006 [199, 240] loss: 0.006 [199, 300] loss: 0.006 [199, 360] loss: 0.006 Epoch: 199 -> Loss: 0.00762056699023 Epoch: 199 -> Test Accuracy: 91.29 [200, 60] loss: 0.006 [200, 120] loss: 0.006 [200, 180] loss: 0.007 [200, 240] loss: 0.006 [200, 300] loss: 0.006 [200, 360] loss: 0.007 Epoch: 200 -> Loss: 0.0176921673119 Epoch: 200 -> Test Accuracy: 91.39 Finished Training
In [11]:
# save variables
fm.save_variable([class_NIN_loss_log, class_NIN_test_accuracy_log], "supervised_NIN")
Semi-supervised Learning¶
In [12]:
# initialize networks
semi_net = fm.load_net("RotNet_rotation_200_4_block_net")
In [13]:
semi_loss_log, semi_accuracy_log, super_loss_log, super_accuracy_log = tr.train_semi([20, 100, 400, 1000, 5000], 10,
trainset, testset, 128, [0.1, 0.02, 0.004, 0.0008], [35, 70, 85, 100], [0.1, 0.02, 0.004, 0.0008],
[60, 120, 160, 200], 0.9, 5e-4, semi_net, criterion)
Epoch: 1 -> Loss: 2.27038741112 Epoch: 1 -> Test Accuracy: 38.2 Epoch: 2 -> Loss: 1.25083875656 Epoch: 2 -> Test Accuracy: 50.25 Epoch: 3 -> Loss: 1.05511724949 Epoch: 3 -> Test Accuracy: 54.67 Epoch: 4 -> Loss: 0.768583297729 Epoch: 4 -> Test Accuracy: 55.78 Epoch: 5 -> Loss: 0.580367803574 Epoch: 5 -> Test Accuracy: 57.47 Epoch: 6 -> Loss: 0.346341788769 Epoch: 6 -> Test Accuracy: 58.57 Epoch: 7 -> Loss: 0.391008704901 Epoch: 7 -> Test Accuracy: 58.75 Epoch: 8 -> Loss: 0.21460570395 Epoch: 8 -> Test Accuracy: 59.13 Epoch: 9 -> Loss: 0.178526058793 Epoch: 9 -> Test Accuracy: 59.64 Epoch: 10 -> Loss: 0.193584755063 Epoch: 10 -> Test Accuracy: 59.73 Epoch: 11 -> Loss: 0.0977763086557 Epoch: 11 -> Test Accuracy: 59.9 Epoch: 12 -> Loss: 0.104682013392 Epoch: 12 -> Test Accuracy: 60.15 Epoch: 13 -> Loss: 0.0651761591434 Epoch: 13 -> Test Accuracy: 60.38 Epoch: 14 -> Loss: 0.0543492771685 Epoch: 14 -> Test Accuracy: 60.51 Epoch: 15 -> Loss: 0.0612868741155 Epoch: 15 -> Test Accuracy: 60.52 Epoch: 16 -> Loss: 0.042023614049 Epoch: 16 -> Test Accuracy: 60.71 Epoch: 17 -> Loss: 0.0215672515333 Epoch: 17 -> Test Accuracy: 61.07 Epoch: 18 -> Loss: 0.0256984103471 Epoch: 18 -> Test Accuracy: 61.27 Epoch: 19 -> Loss: 0.0214232727885 Epoch: 19 -> Test Accuracy: 61.38 Epoch: 20 -> Loss: 0.0318058468401 Epoch: 20 -> Test Accuracy: 61.63 Epoch: 21 -> Loss: 0.0183569863439 Epoch: 21 -> Test Accuracy: 61.75 Epoch: 22 -> Loss: 0.0253742858768 Epoch: 22 -> Test Accuracy: 61.72 Epoch: 23 -> Loss: 0.0246881116182 Epoch: 23 -> Test Accuracy: 61.89 Epoch: 24 -> Loss: 0.017155315727 Epoch: 24 -> Test Accuracy: 61.95 Epoch: 25 -> Loss: 0.0272710844874 Epoch: 25 -> Test Accuracy: 62.08 Epoch: 26 -> Loss: 0.0113559830934 Epoch: 26 -> Test Accuracy: 62.15 Epoch: 27 -> Loss: 0.0115368366241 Epoch: 27 -> Test Accuracy: 62.23 Epoch: 28 -> Loss: 0.00796966440976 Epoch: 28 -> Test Accuracy: 62.23 Epoch: 29 -> Loss: 0.0120957894251 Epoch: 29 -> Test Accuracy: 62.25 Epoch: 30 -> Loss: 0.0114964116365 Epoch: 30 -> Test Accuracy: 62.24 Epoch: 31 -> Loss: 0.0101838968694 Epoch: 31 -> Test Accuracy: 62.28 Epoch: 32 -> Loss: 0.01213702932 Epoch: 32 -> Test Accuracy: 62.43 Epoch: 33 -> Loss: 0.0101211536676 Epoch: 33 -> Test Accuracy: 62.46 Epoch: 34 -> Loss: 0.010621256195 Epoch: 34 -> Test Accuracy: 62.56 Epoch: 35 -> Loss: 0.00744563341141 Epoch: 35 -> Test Accuracy: 62.55 Epoch: 36 -> Loss: 0.0056139761582 Epoch: 36 -> Test Accuracy: 62.55 Epoch: 37 -> Loss: 0.00848679430783 Epoch: 37 -> Test Accuracy: 62.56 Epoch: 38 -> Loss: 0.00818242598325 Epoch: 38 -> Test Accuracy: 62.57 Epoch: 39 -> Loss: 0.00871019251645 Epoch: 39 -> Test Accuracy: 62.55 Epoch: 40 -> Loss: 0.00552598619834 Epoch: 40 -> Test Accuracy: 62.58 Epoch: 41 -> Loss: 0.00760336732492 Epoch: 41 -> Test Accuracy: 62.6 Epoch: 42 -> Loss: 0.00530742947012 Epoch: 42 -> Test Accuracy: 62.61 Epoch: 43 -> Loss: 0.00672916555777 Epoch: 43 -> Test Accuracy: 62.63 Epoch: 44 -> Loss: 0.00878243334591 Epoch: 44 -> Test Accuracy: 62.59 Epoch: 45 -> Loss: 0.00568009726703 Epoch: 45 -> Test Accuracy: 62.62 Epoch: 46 -> Loss: 0.0154432523996 Epoch: 46 -> Test Accuracy: 62.62 Epoch: 47 -> Loss: 0.00832791440189 Epoch: 47 -> Test Accuracy: 62.68 Epoch: 48 -> Loss: 0.00655911350623 Epoch: 48 -> Test Accuracy: 62.73 Epoch: 49 -> Loss: 0.00694914674386 Epoch: 49 -> Test Accuracy: 62.71 Epoch: 50 -> Loss: 0.00769170792773 Epoch: 50 -> Test Accuracy: 62.77 Epoch: 51 -> Loss: 0.00710220448673 Epoch: 51 -> Test Accuracy: 62.73 Epoch: 52 -> Loss: 0.00946570094675 Epoch: 52 -> Test Accuracy: 62.66 Epoch: 53 -> Loss: 0.0101848579943 Epoch: 53 -> Test Accuracy: 62.63 Epoch: 54 -> Loss: 0.0074348449707 Epoch: 54 -> Test Accuracy: 62.65 Epoch: 55 -> Loss: 0.00517143821344 Epoch: 55 -> Test Accuracy: 62.62 Epoch: 56 -> Loss: 0.00575730530545 Epoch: 56 -> Test Accuracy: 62.59 Epoch: 57 -> Loss: 0.00401691580191 Epoch: 57 -> Test Accuracy: 62.6 Epoch: 58 -> Loss: 0.00543261226267 Epoch: 58 -> Test Accuracy: 62.62 Epoch: 59 -> Loss: 0.0074658789672 Epoch: 59 -> Test Accuracy: 62.62 Epoch: 60 -> Loss: 0.00945476070046 Epoch: 60 -> Test Accuracy: 62.6 Epoch: 61 -> Loss: 0.00629272079095 Epoch: 61 -> Test Accuracy: 62.58 Epoch: 62 -> Loss: 0.00855879019946 Epoch: 62 -> Test Accuracy: 62.62 Epoch: 63 -> Loss: 0.00633266242221 Epoch: 63 -> Test Accuracy: 62.62 Epoch: 64 -> Loss: 0.010481165722 Epoch: 64 -> Test Accuracy: 62.62 Epoch: 65 -> Loss: 0.00626732921228 Epoch: 65 -> Test Accuracy: 62.64 Epoch: 66 -> Loss: 0.00693629868329 Epoch: 66 -> Test Accuracy: 62.66 Epoch: 67 -> Loss: 0.00575800053775 Epoch: 67 -> Test Accuracy: 62.76 Epoch: 68 -> Loss: 0.00760897016153 Epoch: 68 -> Test Accuracy: 62.69 Epoch: 69 -> Loss: 0.00582766532898 Epoch: 69 -> Test Accuracy: 62.71 Epoch: 70 -> Loss: 0.00553533760831 Epoch: 70 -> Test Accuracy: 62.71 Epoch: 71 -> Loss: 0.00661115534604 Epoch: 71 -> Test Accuracy: 62.73 Epoch: 72 -> Loss: 0.00559503491968 Epoch: 72 -> Test Accuracy: 62.72 Epoch: 73 -> Loss: 0.00652698008344 Epoch: 73 -> Test Accuracy: 62.72 Epoch: 74 -> Loss: 0.00753558333963 Epoch: 74 -> Test Accuracy: 62.72 Epoch: 75 -> Loss: 0.00685798469931 Epoch: 75 -> Test Accuracy: 62.73 Epoch: 76 -> Loss: 0.00662627490237 Epoch: 76 -> Test Accuracy: 62.71 Epoch: 77 -> Loss: 0.00527127599344 Epoch: 77 -> Test Accuracy: 62.71 Epoch: 78 -> Loss: 0.00652544386685 Epoch: 78 -> Test Accuracy: 62.75 Epoch: 79 -> Loss: 0.00536925252527 Epoch: 79 -> Test Accuracy: 62.73 Epoch: 80 -> Loss: 0.0083562200889 Epoch: 80 -> Test Accuracy: 62.74 Epoch: 81 -> Loss: 0.00584895722568 Epoch: 81 -> Test Accuracy: 62.75 Epoch: 82 -> Loss: 0.0057873330079 Epoch: 82 -> Test Accuracy: 62.76 Epoch: 83 -> Loss: 0.00544299697503 Epoch: 83 -> Test Accuracy: 62.78 Epoch: 84 -> Loss: 0.00688730366528 Epoch: 84 -> Test Accuracy: 62.78 Epoch: 85 -> Loss: 0.00633131805807 Epoch: 85 -> Test Accuracy: 62.79 Epoch: 86 -> Loss: 0.0042668976821 Epoch: 86 -> Test Accuracy: 62.79 Epoch: 87 -> Loss: 0.00515410630032 Epoch: 87 -> Test Accuracy: 62.79 Epoch: 88 -> Loss: 0.00569624360651 Epoch: 88 -> Test Accuracy: 62.79 Epoch: 89 -> Loss: 0.0118077462539 Epoch: 89 -> Test Accuracy: 62.79 Epoch: 90 -> Loss: 0.00667405128479 Epoch: 90 -> Test Accuracy: 62.78 Epoch: 91 -> Loss: 0.00953802838922 Epoch: 91 -> Test Accuracy: 62.78 Epoch: 92 -> Loss: 0.00702718878165 Epoch: 92 -> Test Accuracy: 62.78 Epoch: 93 -> Loss: 0.00599869759753 Epoch: 93 -> Test Accuracy: 62.78 Epoch: 94 -> Loss: 0.00567687861621 Epoch: 94 -> Test Accuracy: 62.78 Epoch: 95 -> Loss: 0.00786071363837 Epoch: 95 -> Test Accuracy: 62.78 Epoch: 96 -> Loss: 0.00724229542539 Epoch: 96 -> Test Accuracy: 62.77 Epoch: 97 -> Loss: 0.00592624489218 Epoch: 97 -> Test Accuracy: 62.76 Epoch: 98 -> Loss: 0.00631262874231 Epoch: 98 -> Test Accuracy: 62.76 Epoch: 99 -> Loss: 0.00812559667975 Epoch: 99 -> Test Accuracy: 62.76 Epoch: 100 -> Loss: 0.00592105928808 Epoch: 100 -> Test Accuracy: 62.76 Finished Training Epoch: 1 -> Loss: 2.65849399567 Epoch: 1 -> Test Accuracy: 17.76 Epoch: 2 -> Loss: 2.65423417091 Epoch: 2 -> Test Accuracy: 19.05 Epoch: 3 -> Loss: 2.54765224457 Epoch: 3 -> Test Accuracy: 21.98 Epoch: 4 -> Loss: 1.95757567883 Epoch: 4 -> Test Accuracy: 23.75 Epoch: 5 -> Loss: 1.69368875027 Epoch: 5 -> Test Accuracy: 25.06 Epoch: 6 -> Loss: 1.53716754913 Epoch: 6 -> Test Accuracy: 26.19 Epoch: 7 -> Loss: 1.46634602547 Epoch: 7 -> Test Accuracy: 28.46 Epoch: 8 -> Loss: 1.33804786205 Epoch: 8 -> Test Accuracy: 29.75 Epoch: 9 -> Loss: 1.22867488861 Epoch: 9 -> Test Accuracy: 29.07 Epoch: 10 -> Loss: 1.29244184494 Epoch: 10 -> Test Accuracy: 30.62 Epoch: 11 -> Loss: 1.11101794243 Epoch: 11 -> Test Accuracy: 29.46 Epoch: 12 -> Loss: 0.977714300156 Epoch: 12 -> Test Accuracy: 30.08 Epoch: 13 -> Loss: 0.839048564434 Epoch: 13 -> Test Accuracy: 30.15 Epoch: 14 -> Loss: 0.728604733944 Epoch: 14 -> Test Accuracy: 30.15 Epoch: 15 -> Loss: 0.652540504932 Epoch: 15 -> Test Accuracy: 28.31 Epoch: 16 -> Loss: 0.60499227047 Epoch: 16 -> Test Accuracy: 29.02 Epoch: 17 -> Loss: 0.671340584755 Epoch: 17 -> Test Accuracy: 29.49 Epoch: 18 -> Loss: 0.610807180405 Epoch: 18 -> Test Accuracy: 31.09 Epoch: 19 -> Loss: 0.481778770685 Epoch: 19 -> Test Accuracy: 30.48 Epoch: 20 -> Loss: 0.695512056351 Epoch: 20 -> Test Accuracy: 28.5 Epoch: 21 -> Loss: 0.49528503418 Epoch: 21 -> Test Accuracy: 29.71 Epoch: 22 -> Loss: 0.508057355881 Epoch: 22 -> Test Accuracy: 29.87 Epoch: 23 -> Loss: 0.484232485294 Epoch: 23 -> Test Accuracy: 30.31 Epoch: 24 -> Loss: 0.409639388323 Epoch: 24 -> Test Accuracy: 29.3 Epoch: 25 -> Loss: 0.400426089764 Epoch: 25 -> Test Accuracy: 30.4 Epoch: 26 -> Loss: 0.366675168276 Epoch: 26 -> Test Accuracy: 31.84 Epoch: 27 -> Loss: 0.213309317827 Epoch: 27 -> Test Accuracy: 29.36 Epoch: 28 -> Loss: 0.190486699343 Epoch: 28 -> Test Accuracy: 30.37 Epoch: 29 -> Loss: 0.183573544025 Epoch: 29 -> Test Accuracy: 30.41 Epoch: 30 -> Loss: 0.121843934059 Epoch: 30 -> Test Accuracy: 30.46 Epoch: 31 -> Loss: 0.0818746984005 Epoch: 31 -> Test Accuracy: 30.43 Epoch: 32 -> Loss: 0.0889547541738 Epoch: 32 -> Test Accuracy: 30.57 Epoch: 33 -> Loss: 0.146276921034 Epoch: 33 -> Test Accuracy: 30.31 Epoch: 34 -> Loss: 0.12046982348 Epoch: 34 -> Test Accuracy: 30.44 Epoch: 35 -> Loss: 0.0818599760532 Epoch: 35 -> Test Accuracy: 29.31 Epoch: 36 -> Loss: 0.093117967248 Epoch: 36 -> Test Accuracy: 31.31 Epoch: 37 -> Loss: 0.108391337097 Epoch: 37 -> Test Accuracy: 31.04 Epoch: 38 -> Loss: 0.225953370333 Epoch: 38 -> Test Accuracy: 30.36 Epoch: 39 -> Loss: 0.0714892223477 Epoch: 39 -> Test Accuracy: 30.66 Epoch: 40 -> Loss: 0.102141693234 Epoch: 40 -> Test Accuracy: 31.18 Epoch: 41 -> Loss: 0.0641302764416 Epoch: 41 -> Test Accuracy: 31.13 Epoch: 42 -> Loss: 0.0632255971432 Epoch: 42 -> Test Accuracy: 31.5 Epoch: 43 -> Loss: 0.058499738574 Epoch: 43 -> Test Accuracy: 31.91 Epoch: 44 -> Loss: 0.0465745925903 Epoch: 44 -> Test Accuracy: 32.22 Epoch: 45 -> Loss: 0.06388682127 Epoch: 45 -> Test Accuracy: 31.56 Epoch: 46 -> Loss: 0.0645600110292 Epoch: 46 -> Test Accuracy: 30.91 Epoch: 47 -> Loss: 0.0624152608216 Epoch: 47 -> Test Accuracy: 30.72 Epoch: 48 -> Loss: 0.0933792591095 Epoch: 48 -> Test Accuracy: 30.18 Epoch: 49 -> Loss: 0.0777014568448 Epoch: 49 -> Test Accuracy: 29.82 Epoch: 50 -> Loss: 0.079465046525 Epoch: 50 -> Test Accuracy: 29.38 Epoch: 51 -> Loss: 0.0368623062968 Epoch: 51 -> Test Accuracy: 30.6 Epoch: 52 -> Loss: 0.0495228022337 Epoch: 52 -> Test Accuracy: 30.79 Epoch: 53 -> Loss: 0.0355994179845 Epoch: 53 -> Test Accuracy: 30.73 Epoch: 54 -> Loss: 0.0513694621623 Epoch: 54 -> Test Accuracy: 30.46 Epoch: 55 -> Loss: 0.0200139954686 Epoch: 55 -> Test Accuracy: 30.79 Epoch: 56 -> Loss: 0.0263384841383 Epoch: 56 -> Test Accuracy: 30.83 Epoch: 57 -> Loss: 0.0260647013783 Epoch: 57 -> Test Accuracy: 30.98 Epoch: 58 -> Loss: 0.013524711132 Epoch: 58 -> Test Accuracy: 31.37 Epoch: 59 -> Loss: 0.0266250167042 Epoch: 59 -> Test Accuracy: 31.43 Epoch: 60 -> Loss: 0.0162335038185 Epoch: 60 -> Test Accuracy: 30.95 Epoch: 61 -> Loss: 0.030623389408 Epoch: 61 -> Test Accuracy: 31.17 Epoch: 62 -> Loss: 0.0176982078701 Epoch: 62 -> Test Accuracy: 31.21 Epoch: 63 -> Loss: 0.0143640972674 Epoch: 63 -> Test Accuracy: 31.37 Epoch: 64 -> Loss: 0.0239562653005 Epoch: 64 -> Test Accuracy: 31.48 Epoch: 65 -> Loss: 0.00683768605813 Epoch: 65 -> Test Accuracy: 31.25 Epoch: 66 -> Loss: 0.0121848322451 Epoch: 66 -> Test Accuracy: 31.28 Epoch: 67 -> Loss: 0.00535400724038 Epoch: 67 -> Test Accuracy: 31.24 Epoch: 68 -> Loss: 0.0268093682826 Epoch: 68 -> Test Accuracy: 31.1 Epoch: 69 -> Loss: 0.0117152733728 Epoch: 69 -> Test Accuracy: 31.16 Epoch: 70 -> Loss: 0.014351089485 Epoch: 70 -> Test Accuracy: 31.23 Epoch: 71 -> Loss: 0.0194191131741 Epoch: 71 -> Test Accuracy: 31.08 Epoch: 72 -> Loss: 0.0089210011065 Epoch: 72 -> Test Accuracy: 31.08 Epoch: 73 -> Loss: 0.00735072977841 Epoch: 73 -> Test Accuracy: 31.08 Epoch: 74 -> Loss: 0.00625147437677 Epoch: 74 -> Test Accuracy: 31.25 Epoch: 75 -> Loss: 0.00578457769006 Epoch: 75 -> Test Accuracy: 31.31 Epoch: 76 -> Loss: 0.00498634576797 Epoch: 76 -> Test Accuracy: 31.34 Epoch: 77 -> Loss: 0.00985836330801 Epoch: 77 -> Test Accuracy: 31.35 Epoch: 78 -> Loss: 0.0651436969638 Epoch: 78 -> Test Accuracy: 31.62 Epoch: 79 -> Loss: 0.0210603680462 Epoch: 79 -> Test Accuracy: 31.77 Epoch: 80 -> Loss: 0.0499172620475 Epoch: 80 -> Test Accuracy: 31.56 Epoch: 81 -> Loss: 0.00692547671497 Epoch: 81 -> Test Accuracy: 31.42 Epoch: 82 -> Loss: 0.00814078934491 Epoch: 82 -> Test Accuracy: 31.31 Epoch: 83 -> Loss: 0.0120227206498 Epoch: 83 -> Test Accuracy: 31.35 Epoch: 84 -> Loss: 0.00890287384391 Epoch: 84 -> Test Accuracy: 31.36 Epoch: 85 -> Loss: 0.00359779596329 Epoch: 85 -> Test Accuracy: 31.45 Epoch: 86 -> Loss: 0.00684562651441 Epoch: 86 -> Test Accuracy: 31.4 Epoch: 87 -> Loss: 0.0163849070668 Epoch: 87 -> Test Accuracy: 31.4 Epoch: 88 -> Loss: 0.00557531928644 Epoch: 88 -> Test Accuracy: 31.25 Epoch: 89 -> Loss: 0.0057560140267 Epoch: 89 -> Test Accuracy: 31.26 Epoch: 90 -> Loss: 0.0173290707171 Epoch: 90 -> Test Accuracy: 31.32 Epoch: 91 -> Loss: 0.00566754071042 Epoch: 91 -> Test Accuracy: 31.4 Epoch: 92 -> Loss: 0.00838792324066 Epoch: 92 -> Test Accuracy: 31.53 Epoch: 93 -> Loss: 0.00608476670459 Epoch: 93 -> Test Accuracy: 31.63 Epoch: 94 -> Loss: 0.00523587083444 Epoch: 94 -> Test Accuracy: 31.67 Epoch: 95 -> Loss: 0.00402182340622 Epoch: 95 -> Test Accuracy: 31.78 Epoch: 96 -> Loss: 0.00507264677435 Epoch: 96 -> Test Accuracy: 31.74 Epoch: 97 -> Loss: 0.012475874275 Epoch: 97 -> Test Accuracy: 31.77 Epoch: 98 -> Loss: 0.0043402579613 Epoch: 98 -> Test Accuracy: 31.85 Epoch: 99 -> Loss: 0.00990788824856 Epoch: 99 -> Test Accuracy: 31.88 Epoch: 100 -> Loss: 0.00575581518933 Epoch: 100 -> Test Accuracy: 31.85 Epoch: 101 -> Loss: 0.0046511227265 Epoch: 101 -> Test Accuracy: 31.94 Epoch: 102 -> Loss: 0.00604559993371 Epoch: 102 -> Test Accuracy: 31.93 Epoch: 103 -> Loss: 0.00441769743338 Epoch: 103 -> Test Accuracy: 31.89 Epoch: 104 -> Loss: 0.0139822624624 Epoch: 104 -> Test Accuracy: 31.81 Epoch: 105 -> Loss: 0.00459103472531 Epoch: 105 -> Test Accuracy: 31.66 Epoch: 106 -> Loss: 0.00472034327686 Epoch: 106 -> Test Accuracy: 31.53 Epoch: 107 -> Loss: 0.00645211013034 Epoch: 107 -> Test Accuracy: 31.45 Epoch: 108 -> Loss: 0.00555882183835 Epoch: 108 -> Test Accuracy: 31.4 Epoch: 109 -> Loss: 0.0086893774569 Epoch: 109 -> Test Accuracy: 31.34 Epoch: 110 -> Loss: 0.00391525682062 Epoch: 110 -> Test Accuracy: 31.29 Epoch: 111 -> Loss: 0.00485305674374 Epoch: 111 -> Test Accuracy: 31.16 Epoch: 112 -> Loss: 0.0193780455738 Epoch: 112 -> Test Accuracy: 31.42 Epoch: 113 -> Loss: 0.00531618483365 Epoch: 113 -> Test Accuracy: 31.51 Epoch: 114 -> Loss: 0.0139916278422 Epoch: 114 -> Test Accuracy: 31.66 Epoch: 115 -> Loss: 0.00473481416702 Epoch: 115 -> Test Accuracy: 31.7 Epoch: 116 -> Loss: 0.00936312135309 Epoch: 116 -> Test Accuracy: 31.6 Epoch: 117 -> Loss: 0.00522234709933 Epoch: 117 -> Test Accuracy: 31.6 Epoch: 118 -> Loss: 0.00459918053821 Epoch: 118 -> Test Accuracy: 31.64 Epoch: 119 -> Loss: 0.0116134220734 Epoch: 119 -> Test Accuracy: 31.65 Epoch: 120 -> Loss: 0.00426397053525 Epoch: 120 -> Test Accuracy: 31.66 Epoch: 121 -> Loss: 0.00583248678595 Epoch: 121 -> Test Accuracy: 31.65 Epoch: 122 -> Loss: 0.00510450219736 Epoch: 122 -> Test Accuracy: 31.68 Epoch: 123 -> Loss: 0.0061276354827 Epoch: 123 -> Test Accuracy: 31.66 Epoch: 124 -> Loss: 0.0043531190604 Epoch: 124 -> Test Accuracy: 31.64 Epoch: 125 -> Loss: 0.00450514443219 Epoch: 125 -> Test Accuracy: 31.64 Epoch: 126 -> Loss: 0.00457972288132 Epoch: 126 -> Test Accuracy: 31.67 Epoch: 127 -> Loss: 0.00374668836594 Epoch: 127 -> Test Accuracy: 31.65 Epoch: 128 -> Loss: 0.00424822187051 Epoch: 128 -> Test Accuracy: 31.67 Epoch: 129 -> Loss: 0.00375601975247 Epoch: 129 -> Test Accuracy: 31.63 Epoch: 130 -> Loss: 0.0044202208519 Epoch: 130 -> Test Accuracy: 31.63 Epoch: 131 -> Loss: 0.0126273762435 Epoch: 131 -> Test Accuracy: 31.57 Epoch: 132 -> Loss: 0.00593539746478 Epoch: 132 -> Test Accuracy: 31.6 Epoch: 133 -> Loss: 0.00738661829382 Epoch: 133 -> Test Accuracy: 31.61 Epoch: 134 -> Loss: 0.00854841899127 Epoch: 134 -> Test Accuracy: 31.63 Epoch: 135 -> Loss: 0.00512178055942 Epoch: 135 -> Test Accuracy: 31.59 Epoch: 136 -> Loss: 0.00753069575876 Epoch: 136 -> Test Accuracy: 31.58 Epoch: 137 -> Loss: 0.00834982283413 Epoch: 137 -> Test Accuracy: 31.63 Epoch: 138 -> Loss: 0.0110670793802 Epoch: 138 -> Test Accuracy: 31.67 Epoch: 139 -> Loss: 0.00301223341376 Epoch: 139 -> Test Accuracy: 31.66 Epoch: 140 -> Loss: 0.00494349002838 Epoch: 140 -> Test Accuracy: 31.69 Epoch: 141 -> Loss: 0.0113283265382 Epoch: 141 -> Test Accuracy: 31.69 Epoch: 142 -> Loss: 0.0196752622724 Epoch: 142 -> Test Accuracy: 31.72 Epoch: 143 -> Loss: 0.00411151535809 Epoch: 143 -> Test Accuracy: 31.7 Epoch: 144 -> Loss: 0.00513248331845 Epoch: 144 -> Test Accuracy: 31.63 Epoch: 145 -> Loss: 0.00871815253049 Epoch: 145 -> Test Accuracy: 31.67 Epoch: 146 -> Loss: 0.00703046703711 Epoch: 146 -> Test Accuracy: 31.62 Epoch: 147 -> Loss: 0.00342211453244 Epoch: 147 -> Test Accuracy: 31.58 Epoch: 148 -> Loss: 0.00384040013887 Epoch: 148 -> Test Accuracy: 31.61 Epoch: 149 -> Loss: 0.00589236291125 Epoch: 149 -> Test Accuracy: 31.64 Epoch: 150 -> Loss: 0.00738715473562 Epoch: 150 -> Test Accuracy: 31.62 Epoch: 151 -> Loss: 0.00435251649469 Epoch: 151 -> Test Accuracy: 31.61 Epoch: 152 -> Loss: 0.00312617095187 Epoch: 152 -> Test Accuracy: 31.65 Epoch: 153 -> Loss: 0.00460541248322 Epoch: 153 -> Test Accuracy: 31.67 Epoch: 154 -> Loss: 0.00909231137484 Epoch: 154 -> Test Accuracy: 31.7 Epoch: 155 -> Loss: 0.00584716256708 Epoch: 155 -> Test Accuracy: 31.7 Epoch: 156 -> Loss: 0.0074987676926 Epoch: 156 -> Test Accuracy: 31.64 Epoch: 157 -> Loss: 0.0039552715607 Epoch: 157 -> Test Accuracy: 31.6 Epoch: 158 -> Loss: 0.00470164744183 Epoch: 158 -> Test Accuracy: 31.57 Epoch: 159 -> Loss: 0.00638218037784 Epoch: 159 -> Test Accuracy: 31.58 Epoch: 160 -> Loss: 0.00485002342612 Epoch: 160 -> Test Accuracy: 31.58 Epoch: 161 -> Loss: 0.00382403540425 Epoch: 161 -> Test Accuracy: 31.57 Epoch: 162 -> Loss: 0.00521539989859 Epoch: 162 -> Test Accuracy: 31.58 Epoch: 163 -> Loss: 0.0062480699271 Epoch: 163 -> Test Accuracy: 31.58 Epoch: 164 -> Loss: 0.0029730333481 Epoch: 164 -> Test Accuracy: 31.59 Epoch: 165 -> Loss: 0.0035161243286 Epoch: 165 -> Test Accuracy: 31.59 Epoch: 166 -> Loss: 0.00549176009372 Epoch: 166 -> Test Accuracy: 31.59 Epoch: 167 -> Loss: 0.00459265056998 Epoch: 167 -> Test Accuracy: 31.6 Epoch: 168 -> Loss: 0.00541173107922 Epoch: 168 -> Test Accuracy: 31.61 Epoch: 169 -> Loss: 0.00387814315036 Epoch: 169 -> Test Accuracy: 31.62 Epoch: 170 -> Loss: 0.00673397397622 Epoch: 170 -> Test Accuracy: 31.63 Epoch: 171 -> Loss: 0.00420149834827 Epoch: 171 -> Test Accuracy: 31.64 Epoch: 172 -> Loss: 0.00555738480762 Epoch: 172 -> Test Accuracy: 31.63 Epoch: 173 -> Loss: 0.00288179190829 Epoch: 173 -> Test Accuracy: 31.62 Epoch: 174 -> Loss: 0.00811217259616 Epoch: 174 -> Test Accuracy: 31.62 Epoch: 175 -> Loss: 0.00299417972565 Epoch: 175 -> Test Accuracy: 31.62 Epoch: 176 -> Loss: 0.00542287016287 Epoch: 176 -> Test Accuracy: 31.61 Epoch: 177 -> Loss: 0.00397737137973 Epoch: 177 -> Test Accuracy: 31.61 Epoch: 178 -> Loss: 0.00658184289932 Epoch: 178 -> Test Accuracy: 31.6 Epoch: 179 -> Loss: 0.00398606061935 Epoch: 179 -> Test Accuracy: 31.6 Epoch: 180 -> Loss: 0.00338594126515 Epoch: 180 -> Test Accuracy: 31.59 Epoch: 181 -> Loss: 0.00717590935528 Epoch: 181 -> Test Accuracy: 31.58 Epoch: 182 -> Loss: 0.00511586666107 Epoch: 182 -> Test Accuracy: 31.59 Epoch: 183 -> Loss: 0.00559444539249 Epoch: 183 -> Test Accuracy: 31.59 Epoch: 184 -> Loss: 0.00833480246365 Epoch: 184 -> Test Accuracy: 31.61 Epoch: 185 -> Loss: 0.00428612343967 Epoch: 185 -> Test Accuracy: 31.61 Epoch: 186 -> Loss: 0.00440870411694 Epoch: 186 -> Test Accuracy: 31.6 Epoch: 187 -> Loss: 0.00514138396829 Epoch: 187 -> Test Accuracy: 31.61 Epoch: 188 -> Loss: 0.00491127697751 Epoch: 188 -> Test Accuracy: 31.6 Epoch: 189 -> Loss: 0.00435345713049 Epoch: 189 -> Test Accuracy: 31.59 Epoch: 190 -> Loss: 0.00580557854846 Epoch: 190 -> Test Accuracy: 31.58 Epoch: 191 -> Loss: 0.00609395233914 Epoch: 191 -> Test Accuracy: 31.58 Epoch: 192 -> Loss: 0.0024350087624 Epoch: 192 -> Test Accuracy: 31.59 Epoch: 193 -> Loss: 0.00519629986957 Epoch: 193 -> Test Accuracy: 31.6 Epoch: 194 -> Loss: 0.00413072761148 Epoch: 194 -> Test Accuracy: 31.59 Epoch: 195 -> Loss: 0.0133555531502 Epoch: 195 -> Test Accuracy: 31.58 Epoch: 196 -> Loss: 0.00494430446997 Epoch: 196 -> Test Accuracy: 31.57 Epoch: 197 -> Loss: 0.00487032858655 Epoch: 197 -> Test Accuracy: 31.56 Epoch: 198 -> Loss: 0.00334726436995 Epoch: 198 -> Test Accuracy: 31.56 Epoch: 199 -> Loss: 0.00479398155585 Epoch: 199 -> Test Accuracy: 31.57 Epoch: 200 -> Loss: 0.00458396784961 Epoch: 200 -> Test Accuracy: 31.57 Finished Training Epoch: 1 -> Loss: 1.16978991032 Epoch: 1 -> Test Accuracy: 59.11 Epoch: 2 -> Loss: 0.818692564964 Epoch: 2 -> Test Accuracy: 64.85 Epoch: 3 -> Loss: 0.515661180019 Epoch: 3 -> Test Accuracy: 67.52 Epoch: 4 -> Loss: 0.498400062323 Epoch: 4 -> Test Accuracy: 69.39 Epoch: 5 -> Loss: 0.295192241669 Epoch: 5 -> Test Accuracy: 69.17 Epoch: 6 -> Loss: 0.346483647823 Epoch: 6 -> Test Accuracy: 70.04 Epoch: 7 -> Loss: 0.35073107481 Epoch: 7 -> Test Accuracy: 69.81 Epoch: 8 -> Loss: 0.164541393518 Epoch: 8 -> Test Accuracy: 70.21 Epoch: 9 -> Loss: 0.136162385345 Epoch: 9 -> Test Accuracy: 70.43 Epoch: 10 -> Loss: 0.197705596685 Epoch: 10 -> Test Accuracy: 69.83 Epoch: 11 -> Loss: 0.161909133196 Epoch: 11 -> Test Accuracy: 71.09 Epoch: 12 -> Loss: 0.128920659423 Epoch: 12 -> Test Accuracy: 70.28 Epoch: 13 -> Loss: 0.141596734524 Epoch: 13 -> Test Accuracy: 69.96 Epoch: 14 -> Loss: 0.0781158283353 Epoch: 14 -> Test Accuracy: 70.14 Epoch: 15 -> Loss: 0.0786408931017 Epoch: 15 -> Test Accuracy: 70.51 Epoch: 16 -> Loss: 0.0920756980777 Epoch: 16 -> Test Accuracy: 70.42 Epoch: 17 -> Loss: 0.0319330617785 Epoch: 17 -> Test Accuracy: 69.73 Epoch: 18 -> Loss: 0.0327526777983 Epoch: 18 -> Test Accuracy: 70.93 Epoch: 19 -> Loss: 0.0480217300355 Epoch: 19 -> Test Accuracy: 71.4 Epoch: 20 -> Loss: 0.0420212186873 Epoch: 20 -> Test Accuracy: 71.05 Epoch: 21 -> Loss: 0.0309940446168 Epoch: 21 -> Test Accuracy: 71.16 Epoch: 22 -> Loss: 0.0445150509477 Epoch: 22 -> Test Accuracy: 71.39 Epoch: 23 -> Loss: 0.023487329483 Epoch: 23 -> Test Accuracy: 70.6 Epoch: 24 -> Loss: 0.0216508414596 Epoch: 24 -> Test Accuracy: 71.47 Epoch: 25 -> Loss: 0.0365014225245 Epoch: 25 -> Test Accuracy: 71.49 Epoch: 26 -> Loss: 0.0115071814507 Epoch: 26 -> Test Accuracy: 70.8 Epoch: 27 -> Loss: 0.0269990563393 Epoch: 27 -> Test Accuracy: 71.03 Epoch: 28 -> Loss: 0.0126689812168 Epoch: 28 -> Test Accuracy: 71.18 Epoch: 29 -> Loss: 0.0354838557541 Epoch: 29 -> Test Accuracy: 71.09 Epoch: 30 -> Loss: 0.02070748806 Epoch: 30 -> Test Accuracy: 71.47 Epoch: 31 -> Loss: 0.0218499582261 Epoch: 31 -> Test Accuracy: 71.19 Epoch: 32 -> Loss: 0.0233822092414 Epoch: 32 -> Test Accuracy: 70.63 Epoch: 33 -> Loss: 0.0201046839356 Epoch: 33 -> Test Accuracy: 71.1 Epoch: 34 -> Loss: 0.0154700558633 Epoch: 34 -> Test Accuracy: 71.22 Epoch: 35 -> Loss: 0.0124344872311 Epoch: 35 -> Test Accuracy: 71.47 Epoch: 36 -> Loss: 0.015577564016 Epoch: 36 -> Test Accuracy: 71.46 Epoch: 37 -> Loss: 0.0123796053231 Epoch: 37 -> Test Accuracy: 71.68 Epoch: 38 -> Loss: 0.013300373219 Epoch: 38 -> Test Accuracy: 71.61 Epoch: 39 -> Loss: 0.0090951602906 Epoch: 39 -> Test Accuracy: 71.66 Epoch: 40 -> Loss: 0.0186039805412 Epoch: 40 -> Test Accuracy: 71.73 Epoch: 41 -> Loss: 0.00844493694603 Epoch: 41 -> Test Accuracy: 71.96 Epoch: 42 -> Loss: 0.00948825664818 Epoch: 42 -> Test Accuracy: 71.94 Epoch: 43 -> Loss: 0.00697893369943 Epoch: 43 -> Test Accuracy: 71.9 Epoch: 44 -> Loss: 0.00895012356341 Epoch: 44 -> Test Accuracy: 71.79 Epoch: 45 -> Loss: 0.00774023169652 Epoch: 45 -> Test Accuracy: 71.84 Epoch: 46 -> Loss: 0.00836301781237 Epoch: 46 -> Test Accuracy: 71.84 Epoch: 47 -> Loss: 0.0102208806202 Epoch: 47 -> Test Accuracy: 71.81 Epoch: 48 -> Loss: 0.00655163265765 Epoch: 48 -> Test Accuracy: 71.85 Epoch: 49 -> Loss: 0.0093550728634 Epoch: 49 -> Test Accuracy: 71.73 Epoch: 50 -> Loss: 0.00735352607444 Epoch: 50 -> Test Accuracy: 71.74 Epoch: 51 -> Loss: 0.0107001615688 Epoch: 51 -> Test Accuracy: 71.73 Epoch: 52 -> Loss: 0.00513489451259 Epoch: 52 -> Test Accuracy: 71.81 Epoch: 53 -> Loss: 0.018765360117 Epoch: 53 -> Test Accuracy: 71.87 Epoch: 54 -> Loss: 0.012613048777 Epoch: 54 -> Test Accuracy: 71.83 Epoch: 55 -> Loss: 0.0124139096588 Epoch: 55 -> Test Accuracy: 71.87 Epoch: 56 -> Loss: 0.0081590320915 Epoch: 56 -> Test Accuracy: 71.87 Epoch: 57 -> Loss: 0.00564740272239 Epoch: 57 -> Test Accuracy: 71.85 Epoch: 58 -> Loss: 0.00776688428596 Epoch: 58 -> Test Accuracy: 71.85 Epoch: 59 -> Loss: 0.00804035924375 Epoch: 59 -> Test Accuracy: 71.88 Epoch: 60 -> Loss: 0.00311387493275 Epoch: 60 -> Test Accuracy: 71.94 Epoch: 61 -> Loss: 0.00572899216786 Epoch: 61 -> Test Accuracy: 71.97 Epoch: 62 -> Loss: 0.0060246726498 Epoch: 62 -> Test Accuracy: 71.96 Epoch: 63 -> Loss: 0.00840062834322 Epoch: 63 -> Test Accuracy: 71.86 Epoch: 64 -> Loss: 0.00872503314167 Epoch: 64 -> Test Accuracy: 71.85 Epoch: 65 -> Loss: 0.00926504191011 Epoch: 65 -> Test Accuracy: 71.82 Epoch: 66 -> Loss: 0.00849655456841 Epoch: 66 -> Test Accuracy: 71.89 Epoch: 67 -> Loss: 0.00573908817023 Epoch: 67 -> Test Accuracy: 71.89 Epoch: 68 -> Loss: 0.00538792507723 Epoch: 68 -> Test Accuracy: 71.85 Epoch: 69 -> Loss: 0.005356724374 Epoch: 69 -> Test Accuracy: 71.9 Epoch: 70 -> Loss: 0.00815541017801 Epoch: 70 -> Test Accuracy: 71.84 Epoch: 71 -> Loss: 0.00879112072289 Epoch: 71 -> Test Accuracy: 71.86 Epoch: 72 -> Loss: 0.00691547291353 Epoch: 72 -> Test Accuracy: 71.88 Epoch: 73 -> Loss: 0.00516853434965 Epoch: 73 -> Test Accuracy: 71.91 Epoch: 74 -> Loss: 0.00770607823506 Epoch: 74 -> Test Accuracy: 71.88 Epoch: 75 -> Loss: 0.00592618249357 Epoch: 75 -> Test Accuracy: 71.91 Epoch: 76 -> Loss: 0.00764277810231 Epoch: 76 -> Test Accuracy: 71.88 Epoch: 77 -> Loss: 0.00649404991418 Epoch: 77 -> Test Accuracy: 71.91 Epoch: 78 -> Loss: 0.00508999358863 Epoch: 78 -> Test Accuracy: 71.91 Epoch: 79 -> Loss: 0.00620726915076 Epoch: 79 -> Test Accuracy: 71.9 Epoch: 80 -> Loss: 0.00717619759962 Epoch: 80 -> Test Accuracy: 71.86 Epoch: 81 -> Loss: 0.00679866177961 Epoch: 81 -> Test Accuracy: 71.82 Epoch: 82 -> Loss: 0.00567138195038 Epoch: 82 -> Test Accuracy: 71.82 Epoch: 83 -> Loss: 0.00546513171867 Epoch: 83 -> Test Accuracy: 71.84 Epoch: 84 -> Loss: 0.00826396420598 Epoch: 84 -> Test Accuracy: 71.79 Epoch: 85 -> Loss: 0.00735905067995 Epoch: 85 -> Test Accuracy: 71.81 Epoch: 86 -> Loss: 0.00670008454472 Epoch: 86 -> Test Accuracy: 71.82 Epoch: 87 -> Loss: 0.00763354869559 Epoch: 87 -> Test Accuracy: 71.82 Epoch: 88 -> Loss: 0.0086984038353 Epoch: 88 -> Test Accuracy: 71.83 Epoch: 89 -> Loss: 0.0146390432492 Epoch: 89 -> Test Accuracy: 71.83 Epoch: 90 -> Loss: 0.00588022731245 Epoch: 90 -> Test Accuracy: 71.83 Epoch: 91 -> Loss: 0.00926686264575 Epoch: 91 -> Test Accuracy: 71.82 Epoch: 92 -> Loss: 0.00477366242558 Epoch: 92 -> Test Accuracy: 71.81 Epoch: 93 -> Loss: 0.0100494027138 Epoch: 93 -> Test Accuracy: 71.79 Epoch: 94 -> Loss: 0.00503945350647 Epoch: 94 -> Test Accuracy: 71.79 Epoch: 95 -> Loss: 0.00737430946901 Epoch: 95 -> Test Accuracy: 71.79 Epoch: 96 -> Loss: 0.00864776782691 Epoch: 96 -> Test Accuracy: 71.79 Epoch: 97 -> Loss: 0.00645641656592 Epoch: 97 -> Test Accuracy: 71.8 Epoch: 98 -> Loss: 0.00669449102134 Epoch: 98 -> Test Accuracy: 71.81 Epoch: 99 -> Loss: 0.00832581054419 Epoch: 99 -> Test Accuracy: 71.82 Epoch: 100 -> Loss: 0.00613514287397 Epoch: 100 -> Test Accuracy: 71.78 Finished Training Epoch: 1 -> Loss: 2.05860495567 Epoch: 1 -> Test Accuracy: 25.15 Epoch: 2 -> Loss: 1.78287065029 Epoch: 2 -> Test Accuracy: 31.69 Epoch: 3 -> Loss: 1.66563344002 Epoch: 3 -> Test Accuracy: 34.04 Epoch: 4 -> Loss: 1.66787862778 Epoch: 4 -> Test Accuracy: 35.87 Epoch: 5 -> Loss: 1.75792908669 Epoch: 5 -> Test Accuracy: 36.5 Epoch: 6 -> Loss: 1.48490762711 Epoch: 6 -> Test Accuracy: 39.22 Epoch: 7 -> Loss: 1.53739726543 Epoch: 7 -> Test Accuracy: 39.08 Epoch: 8 -> Loss: 1.43894505501 Epoch: 8 -> Test Accuracy: 37.4 Epoch: 9 -> Loss: 1.30448019505 Epoch: 9 -> Test Accuracy: 40.94 Epoch: 10 -> Loss: 1.24514913559 Epoch: 10 -> Test Accuracy: 41.34 Epoch: 11 -> Loss: 1.22006881237 Epoch: 11 -> Test Accuracy: 40.87 Epoch: 12 -> Loss: 1.12812423706 Epoch: 12 -> Test Accuracy: 39.79 Epoch: 13 -> Loss: 1.05018496513 Epoch: 13 -> Test Accuracy: 39.84 Epoch: 14 -> Loss: 0.94600135088 Epoch: 14 -> Test Accuracy: 39.32 Epoch: 15 -> Loss: 0.998095750809 Epoch: 15 -> Test Accuracy: 41.34 Epoch: 16 -> Loss: 0.936528921127 Epoch: 16 -> Test Accuracy: 42.1 Epoch: 17 -> Loss: 0.833999037743 Epoch: 17 -> Test Accuracy: 42.84 Epoch: 18 -> Loss: 1.10956180096 Epoch: 18 -> Test Accuracy: 39.31 Epoch: 19 -> Loss: 1.00225949287 Epoch: 19 -> Test Accuracy: 41.07 Epoch: 20 -> Loss: 0.925218343735 Epoch: 20 -> Test Accuracy: 40.87 Epoch: 21 -> Loss: 0.835498392582 Epoch: 21 -> Test Accuracy: 39.0 Epoch: 22 -> Loss: 0.811240792274 Epoch: 22 -> Test Accuracy: 41.49 Epoch: 23 -> Loss: 0.677923440933 Epoch: 23 -> Test Accuracy: 42.51 Epoch: 24 -> Loss: 0.590499579906 Epoch: 24 -> Test Accuracy: 41.46 Epoch: 25 -> Loss: 0.656732559204 Epoch: 25 -> Test Accuracy: 42.96 Epoch: 26 -> Loss: 0.519568741322 Epoch: 26 -> Test Accuracy: 41.32 Epoch: 27 -> Loss: 0.569069862366 Epoch: 27 -> Test Accuracy: 42.16 Epoch: 28 -> Loss: 0.688864827156 Epoch: 28 -> Test Accuracy: 42.76 Epoch: 29 -> Loss: 0.537516713142 Epoch: 29 -> Test Accuracy: 41.28 Epoch: 30 -> Loss: 0.490972310305 Epoch: 30 -> Test Accuracy: 42.4 Epoch: 31 -> Loss: 0.443497300148 Epoch: 31 -> Test Accuracy: 39.4 Epoch: 32 -> Loss: 0.480938374996 Epoch: 32 -> Test Accuracy: 41.85 Epoch: 33 -> Loss: 0.395873755217 Epoch: 33 -> Test Accuracy: 41.77 Epoch: 34 -> Loss: 0.296258032322 Epoch: 34 -> Test Accuracy: 43.45 Epoch: 35 -> Loss: 0.30421307683 Epoch: 35 -> Test Accuracy: 42.5 Epoch: 36 -> Loss: 0.182013511658 Epoch: 36 -> Test Accuracy: 43.09 Epoch: 37 -> Loss: 0.299805849791 Epoch: 37 -> Test Accuracy: 42.3 Epoch: 38 -> Loss: 0.338209331036 Epoch: 38 -> Test Accuracy: 42.69 Epoch: 39 -> Loss: 0.215267896652 Epoch: 39 -> Test Accuracy: 42.66 Epoch: 40 -> Loss: 0.149372398853 Epoch: 40 -> Test Accuracy: 43.64 Epoch: 41 -> Loss: 0.0939838811755 Epoch: 41 -> Test Accuracy: 43.16 Epoch: 42 -> Loss: 0.212134540081 Epoch: 42 -> Test Accuracy: 42.59 Epoch: 43 -> Loss: 0.258519947529 Epoch: 43 -> Test Accuracy: 43.65 Epoch: 44 -> Loss: 0.104442670941 Epoch: 44 -> Test Accuracy: 43.66 Epoch: 45 -> Loss: 0.179340541363 Epoch: 45 -> Test Accuracy: 44.88 Epoch: 46 -> Loss: 0.251038402319 Epoch: 46 -> Test Accuracy: 44.42 Epoch: 47 -> Loss: 0.157367676497 Epoch: 47 -> Test Accuracy: 43.56 Epoch: 48 -> Loss: 0.168719470501 Epoch: 48 -> Test Accuracy: 44.92 Epoch: 49 -> Loss: 0.167285218835 Epoch: 49 -> Test Accuracy: 43.99 Epoch: 50 -> Loss: 0.0926831290126 Epoch: 50 -> Test Accuracy: 44.11 Epoch: 51 -> Loss: 0.137827664614 Epoch: 51 -> Test Accuracy: 43.54 Epoch: 52 -> Loss: 0.240762472153 Epoch: 52 -> Test Accuracy: 41.44 Epoch: 53 -> Loss: 0.144982472062 Epoch: 53 -> Test Accuracy: 43.09 Epoch: 54 -> Loss: 0.18570753932 Epoch: 54 -> Test Accuracy: 42.68 Epoch: 55 -> Loss: 0.192132502794 Epoch: 55 -> Test Accuracy: 43.06 Epoch: 56 -> Loss: 0.258671820164 Epoch: 56 -> Test Accuracy: 42.77 Epoch: 57 -> Loss: 0.30471265316 Epoch: 57 -> Test Accuracy: 42.62 Epoch: 58 -> Loss: 0.239147558808 Epoch: 58 -> Test Accuracy: 44.32 Epoch: 59 -> Loss: 0.123470336199 Epoch: 59 -> Test Accuracy: 42.78 Epoch: 60 -> Loss: 0.101262368262 Epoch: 60 -> Test Accuracy: 44.99 Epoch: 61 -> Loss: 0.0398100018501 Epoch: 61 -> Test Accuracy: 46.09 Epoch: 62 -> Loss: 0.0573702082038 Epoch: 62 -> Test Accuracy: 46.69 Epoch: 63 -> Loss: 0.0128948185593 Epoch: 63 -> Test Accuracy: 46.87 Epoch: 64 -> Loss: 0.0159148797393 Epoch: 64 -> Test Accuracy: 46.61 Epoch: 65 -> Loss: 0.0188232883811 Epoch: 65 -> Test Accuracy: 47.03 Epoch: 66 -> Loss: 0.00894800480455 Epoch: 66 -> Test Accuracy: 47.06 Epoch: 67 -> Loss: 0.0125405974686 Epoch: 67 -> Test Accuracy: 46.83 Epoch: 68 -> Loss: 0.0152934240177 Epoch: 68 -> Test Accuracy: 46.97 Epoch: 69 -> Loss: 0.0110622961074 Epoch: 69 -> Test Accuracy: 47.0 Epoch: 70 -> Loss: 0.0274295527488 Epoch: 70 -> Test Accuracy: 47.0 Epoch: 71 -> Loss: 0.0108768204227 Epoch: 71 -> Test Accuracy: 47.14 Epoch: 72 -> Loss: 0.0234214700758 Epoch: 72 -> Test Accuracy: 47.18 Epoch: 73 -> Loss: 0.0140837933868 Epoch: 73 -> Test Accuracy: 46.94 Epoch: 74 -> Loss: 0.0126457260922 Epoch: 74 -> Test Accuracy: 46.98 Epoch: 75 -> Loss: 0.00654622679576 Epoch: 75 -> Test Accuracy: 47.04 Epoch: 76 -> Loss: 0.020926296711 Epoch: 76 -> Test Accuracy: 47.03 Epoch: 77 -> Loss: 0.00728975329548 Epoch: 77 -> Test Accuracy: 47.06 Epoch: 78 -> Loss: 0.0109080672264 Epoch: 78 -> Test Accuracy: 46.94 Epoch: 79 -> Loss: 0.00699828239158 Epoch: 79 -> Test Accuracy: 46.85 Epoch: 80 -> Loss: 0.00645909877494 Epoch: 80 -> Test Accuracy: 46.87 Epoch: 81 -> Loss: 0.0141588449478 Epoch: 81 -> Test Accuracy: 47.0 Epoch: 82 -> Loss: 0.00822214409709 Epoch: 82 -> Test Accuracy: 47.06 Epoch: 83 -> Loss: 0.0105800908059 Epoch: 83 -> Test Accuracy: 46.97 Epoch: 84 -> Loss: 0.00409800745547 Epoch: 84 -> Test Accuracy: 46.82 Epoch: 85 -> Loss: 0.00754146371037 Epoch: 85 -> Test Accuracy: 46.87 Epoch: 86 -> Loss: 0.00601873034611 Epoch: 86 -> Test Accuracy: 47.0 Epoch: 87 -> Loss: 0.00431439979002 Epoch: 87 -> Test Accuracy: 46.96 Epoch: 88 -> Loss: 0.00561239616945 Epoch: 88 -> Test Accuracy: 46.82 Epoch: 89 -> Loss: 0.00743468012661 Epoch: 89 -> Test Accuracy: 46.87 Epoch: 90 -> Loss: 0.00627677235752 Epoch: 90 -> Test Accuracy: 46.92 Epoch: 91 -> Loss: 0.01095334813 Epoch: 91 -> Test Accuracy: 46.93 Epoch: 92 -> Loss: 0.00718308892101 Epoch: 92 -> Test Accuracy: 46.93 Epoch: 93 -> Loss: 0.00643192324787 Epoch: 93 -> Test Accuracy: 46.97 Epoch: 94 -> Loss: 0.00616349605843 Epoch: 94 -> Test Accuracy: 47.14 Epoch: 95 -> Loss: 0.00576082104817 Epoch: 95 -> Test Accuracy: 47.17 Epoch: 96 -> Loss: 0.00719059910625 Epoch: 96 -> Test Accuracy: 47.24 Epoch: 97 -> Loss: 0.0111631155014 Epoch: 97 -> Test Accuracy: 47.14 Epoch: 98 -> Loss: 0.00323654594831 Epoch: 98 -> Test Accuracy: 47.26 Epoch: 99 -> Loss: 0.00591270299628 Epoch: 99 -> Test Accuracy: 47.12 Epoch: 100 -> Loss: 0.00590293202549 Epoch: 100 -> Test Accuracy: 47.13 Epoch: 101 -> Loss: 0.0117588918656 Epoch: 101 -> Test Accuracy: 46.9 Epoch: 102 -> Loss: 0.00545776356012 Epoch: 102 -> Test Accuracy: 47.09 Epoch: 103 -> Loss: 0.00437755323946 Epoch: 103 -> Test Accuracy: 46.95 Epoch: 104 -> Loss: 0.00414527859539 Epoch: 104 -> Test Accuracy: 46.93 Epoch: 105 -> Loss: 0.00360571872443 Epoch: 105 -> Test Accuracy: 46.87 Epoch: 106 -> Loss: 0.00386307784356 Epoch: 106 -> Test Accuracy: 46.84 Epoch: 107 -> Loss: 0.00497916107997 Epoch: 107 -> Test Accuracy: 46.9 Epoch: 108 -> Loss: 0.0105075147003 Epoch: 108 -> Test Accuracy: 46.98 Epoch: 109 -> Loss: 0.004586242605 Epoch: 109 -> Test Accuracy: 47.06 Epoch: 110 -> Loss: 0.0039576520212 Epoch: 110 -> Test Accuracy: 47.13 Epoch: 111 -> Loss: 0.00331057491712 Epoch: 111 -> Test Accuracy: 47.13 Epoch: 112 -> Loss: 0.00532398326322 Epoch: 112 -> Test Accuracy: 47.12 Epoch: 113 -> Loss: 0.00696651311591 Epoch: 113 -> Test Accuracy: 47.12 Epoch: 114 -> Loss: 0.00535412039608 Epoch: 114 -> Test Accuracy: 47.23 Epoch: 115 -> Loss: 0.00419379677624 Epoch: 115 -> Test Accuracy: 47.04 Epoch: 116 -> Loss: 0.00531852245331 Epoch: 116 -> Test Accuracy: 47.08 Epoch: 117 -> Loss: 0.0045576277189 Epoch: 117 -> Test Accuracy: 47.05 Epoch: 118 -> Loss: 0.0185530818999 Epoch: 118 -> Test Accuracy: 47.16 Epoch: 119 -> Loss: 0.00319520779885 Epoch: 119 -> Test Accuracy: 47.24 Epoch: 120 -> Loss: 0.00451296102256 Epoch: 120 -> Test Accuracy: 47.24 Epoch: 121 -> Loss: 0.00657169613987 Epoch: 121 -> Test Accuracy: 47.2 Epoch: 122 -> Loss: 0.00609574420378 Epoch: 122 -> Test Accuracy: 47.14 Epoch: 123 -> Loss: 0.00424097152427 Epoch: 123 -> Test Accuracy: 47.2 Epoch: 124 -> Loss: 0.00957666896284 Epoch: 124 -> Test Accuracy: 47.2 Epoch: 125 -> Loss: 0.00429891142994 Epoch: 125 -> Test Accuracy: 47.23 Epoch: 126 -> Loss: 0.00412431592122 Epoch: 126 -> Test Accuracy: 47.18 Epoch: 127 -> Loss: 0.00693973666057 Epoch: 127 -> Test Accuracy: 47.19 Epoch: 128 -> Loss: 0.0054598543793 Epoch: 128 -> Test Accuracy: 47.2 Epoch: 129 -> Loss: 0.00380060309544 Epoch: 129 -> Test Accuracy: 47.25 Epoch: 130 -> Loss: 0.00369319552556 Epoch: 130 -> Test Accuracy: 47.27 Epoch: 131 -> Loss: 0.0040620197542 Epoch: 131 -> Test Accuracy: 47.29 Epoch: 132 -> Loss: 0.00732275145128 Epoch: 132 -> Test Accuracy: 47.27 Epoch: 133 -> Loss: 0.00532024633139 Epoch: 133 -> Test Accuracy: 47.29 Epoch: 134 -> Loss: 0.00293260347098 Epoch: 134 -> Test Accuracy: 47.28 Epoch: 135 -> Loss: 0.00995481014252 Epoch: 135 -> Test Accuracy: 47.24 Epoch: 136 -> Loss: 0.00302057992667 Epoch: 136 -> Test Accuracy: 47.22 Epoch: 137 -> Loss: 0.00484318006784 Epoch: 137 -> Test Accuracy: 47.22 Epoch: 138 -> Loss: 0.00447314046323 Epoch: 138 -> Test Accuracy: 47.18 Epoch: 139 -> Loss: 0.00498041743413 Epoch: 139 -> Test Accuracy: 47.11 Epoch: 140 -> Loss: 0.00438820384443 Epoch: 140 -> Test Accuracy: 47.16 Epoch: 141 -> Loss: 0.00395818380639 Epoch: 141 -> Test Accuracy: 47.22 Epoch: 142 -> Loss: 0.00632557505742 Epoch: 142 -> Test Accuracy: 47.21 Epoch: 143 -> Loss: 0.00284797861241 Epoch: 143 -> Test Accuracy: 47.2 Epoch: 144 -> Loss: 0.00460153352469 Epoch: 144 -> Test Accuracy: 47.2 Epoch: 145 -> Loss: 0.00562093826011 Epoch: 145 -> Test Accuracy: 47.19 Epoch: 146 -> Loss: 0.00456475745887 Epoch: 146 -> Test Accuracy: 47.19 Epoch: 147 -> Loss: 0.00395743642002 Epoch: 147 -> Test Accuracy: 47.17 Epoch: 148 -> Loss: 0.00426904950291 Epoch: 148 -> Test Accuracy: 47.19 Epoch: 149 -> Loss: 0.00532081490383 Epoch: 149 -> Test Accuracy: 47.17 Epoch: 150 -> Loss: 0.00454872380942 Epoch: 150 -> Test Accuracy: 47.22 Epoch: 151 -> Loss: 0.00441893702373 Epoch: 151 -> Test Accuracy: 47.25 Epoch: 152 -> Loss: 0.00605936255306 Epoch: 152 -> Test Accuracy: 47.22 Epoch: 153 -> Loss: 0.00341761577874 Epoch: 153 -> Test Accuracy: 47.21 Epoch: 154 -> Loss: 0.00391801958904 Epoch: 154 -> Test Accuracy: 47.16 Epoch: 155 -> Loss: 0.00770230498165 Epoch: 155 -> Test Accuracy: 47.17 Epoch: 156 -> Loss: 0.00276749418117 Epoch: 156 -> Test Accuracy: 47.12 Epoch: 157 -> Loss: 0.00425694091246 Epoch: 157 -> Test Accuracy: 47.13 Epoch: 158 -> Loss: 0.00491246813908 Epoch: 158 -> Test Accuracy: 47.11 Epoch: 159 -> Loss: 0.0060608247295 Epoch: 159 -> Test Accuracy: 47.11 Epoch: 160 -> Loss: 0.00407360168174 Epoch: 160 -> Test Accuracy: 47.16 Epoch: 161 -> Loss: 0.00385200069286 Epoch: 161 -> Test Accuracy: 47.16 Epoch: 162 -> Loss: 0.00673266546801 Epoch: 162 -> Test Accuracy: 47.16 Epoch: 163 -> Loss: 0.00338697899133 Epoch: 163 -> Test Accuracy: 47.15 Epoch: 164 -> Loss: 0.00463841063902 Epoch: 164 -> Test Accuracy: 47.14 Epoch: 165 -> Loss: 0.0174027401954 Epoch: 165 -> Test Accuracy: 47.14 Epoch: 166 -> Loss: 0.0067554069683 Epoch: 166 -> Test Accuracy: 47.14 Epoch: 167 -> Loss: 0.00396752357483 Epoch: 167 -> Test Accuracy: 47.19 Epoch: 168 -> Loss: 0.00519067980349 Epoch: 168 -> Test Accuracy: 47.21 Epoch: 169 -> Loss: 0.00537413358688 Epoch: 169 -> Test Accuracy: 47.2 Epoch: 170 -> Loss: 0.00553401606157 Epoch: 170 -> Test Accuracy: 47.19 Epoch: 171 -> Loss: 0.00237574032508 Epoch: 171 -> Test Accuracy: 47.17 Epoch: 172 -> Loss: 0.00498402584344 Epoch: 172 -> Test Accuracy: 47.17 Epoch: 173 -> Loss: 0.00355258351192 Epoch: 173 -> Test Accuracy: 47.16 Epoch: 174 -> Loss: 0.00436259247363 Epoch: 174 -> Test Accuracy: 47.17 Epoch: 175 -> Loss: 0.0060517648235 Epoch: 175 -> Test Accuracy: 47.19 Epoch: 176 -> Loss: 0.00835307780653 Epoch: 176 -> Test Accuracy: 47.19 Epoch: 177 -> Loss: 0.00520370109007 Epoch: 177 -> Test Accuracy: 47.19 Epoch: 178 -> Loss: 0.00470145372674 Epoch: 178 -> Test Accuracy: 47.18 Epoch: 179 -> Loss: 0.00333922193386 Epoch: 179 -> Test Accuracy: 47.17 Epoch: 180 -> Loss: 0.00964744295925 Epoch: 180 -> Test Accuracy: 47.15 Epoch: 181 -> Loss: 0.00630944035947 Epoch: 181 -> Test Accuracy: 47.13 Epoch: 182 -> Loss: 0.00370637280867 Epoch: 182 -> Test Accuracy: 47.13 Epoch: 183 -> Loss: 0.00374528532848 Epoch: 183 -> Test Accuracy: 47.1 Epoch: 184 -> Loss: 0.00446284748614 Epoch: 184 -> Test Accuracy: 47.11 Epoch: 185 -> Loss: 0.00385080859996 Epoch: 185 -> Test Accuracy: 47.1 Epoch: 186 -> Loss: 0.00440237624571 Epoch: 186 -> Test Accuracy: 47.1 Epoch: 187 -> Loss: 0.00561613310128 Epoch: 187 -> Test Accuracy: 47.1 Epoch: 188 -> Loss: 0.00740049453452 Epoch: 188 -> Test Accuracy: 47.1 Epoch: 189 -> Loss: 0.00424044858664 Epoch: 189 -> Test Accuracy: 47.11 Epoch: 190 -> Loss: 0.00532710086554 Epoch: 190 -> Test Accuracy: 47.09 Epoch: 191 -> Loss: 0.0054624080658 Epoch: 191 -> Test Accuracy: 47.1 Epoch: 192 -> Loss: 0.00476060016081 Epoch: 192 -> Test Accuracy: 47.1 Epoch: 193 -> Loss: 0.00857401359826 Epoch: 193 -> Test Accuracy: 47.13 Epoch: 194 -> Loss: 0.00691185984761 Epoch: 194 -> Test Accuracy: 47.11 Epoch: 195 -> Loss: 0.00707859732211 Epoch: 195 -> Test Accuracy: 47.11 Epoch: 196 -> Loss: 0.00377793493681 Epoch: 196 -> Test Accuracy: 47.08 Epoch: 197 -> Loss: 0.00461611384526 Epoch: 197 -> Test Accuracy: 47.09 Epoch: 198 -> Loss: 0.00951327290386 Epoch: 198 -> Test Accuracy: 47.09 Epoch: 199 -> Loss: 0.004498898983 Epoch: 199 -> Test Accuracy: 47.1 Epoch: 200 -> Loss: 0.00615973677486 Epoch: 200 -> Test Accuracy: 47.1 Finished Training Epoch: 1 -> Loss: 0.802093267441 Epoch: 1 -> Test Accuracy: 66.99 Epoch: 2 -> Loss: 0.424829810858 Epoch: 2 -> Test Accuracy: 74.03 Epoch: 3 -> Loss: 0.504541635513 Epoch: 3 -> Test Accuracy: 75.08 Epoch: 4 -> Loss: 0.546547710896 Epoch: 4 -> Test Accuracy: 75.39 Epoch: 5 -> Loss: 0.378071367741 Epoch: 5 -> Test Accuracy: 74.59 Epoch: 6 -> Loss: 0.171422198415 Epoch: 6 -> Test Accuracy: 75.79 Epoch: 7 -> Loss: 0.351442366838 Epoch: 7 -> Test Accuracy: 76.77 Epoch: 8 -> Loss: 0.239226520061 Epoch: 8 -> Test Accuracy: 76.55 Epoch: 9 -> Loss: 0.245262667537 Epoch: 9 -> Test Accuracy: 77.22 Epoch: 10 -> Loss: 0.276399612427 Epoch: 10 -> Test Accuracy: 76.33 Epoch: 11 -> Loss: 0.186288997531 Epoch: 11 -> Test Accuracy: 77.66 Epoch: 12 -> Loss: 0.0832670927048 Epoch: 12 -> Test Accuracy: 77.25 Epoch: 13 -> Loss: 0.237387895584 Epoch: 13 -> Test Accuracy: 76.98 Epoch: 14 -> Loss: 0.196971029043 Epoch: 14 -> Test Accuracy: 77.13 Epoch: 15 -> Loss: 0.103413499892 Epoch: 15 -> Test Accuracy: 76.8 Epoch: 16 -> Loss: 0.0992646813393 Epoch: 16 -> Test Accuracy: 76.57 Epoch: 17 -> Loss: 0.226926401258 Epoch: 17 -> Test Accuracy: 77.19 Epoch: 18 -> Loss: 0.375478118658 Epoch: 18 -> Test Accuracy: 77.9 Epoch: 19 -> Loss: 0.0377269536257 Epoch: 19 -> Test Accuracy: 77.06 Epoch: 20 -> Loss: 0.0722538083792 Epoch: 20 -> Test Accuracy: 77.72 Epoch: 21 -> Loss: 0.0992808267474 Epoch: 21 -> Test Accuracy: 78.41 Epoch: 22 -> Loss: 0.0456341058016 Epoch: 22 -> Test Accuracy: 77.95 Epoch: 23 -> Loss: 0.0169827789068 Epoch: 23 -> Test Accuracy: 78.21 Epoch: 24 -> Loss: 0.155332386494 Epoch: 24 -> Test Accuracy: 77.87 Epoch: 25 -> Loss: 0.0860455930233 Epoch: 25 -> Test Accuracy: 78.49 Epoch: 26 -> Loss: 0.242450863123 Epoch: 26 -> Test Accuracy: 77.65 Epoch: 27 -> Loss: 0.0242157131433 Epoch: 27 -> Test Accuracy: 77.42 Epoch: 28 -> Loss: 0.216665014625 Epoch: 28 -> Test Accuracy: 78.57 Epoch: 29 -> Loss: 0.0592583715916 Epoch: 29 -> Test Accuracy: 78.14 Epoch: 30 -> Loss: 0.146587640047 Epoch: 30 -> Test Accuracy: 77.86 Epoch: 31 -> Loss: 0.212334707379 Epoch: 31 -> Test Accuracy: 78.53 Epoch: 32 -> Loss: 0.0875750482082 Epoch: 32 -> Test Accuracy: 77.61 Epoch: 33 -> Loss: 0.0500359088182 Epoch: 33 -> Test Accuracy: 77.65 Epoch: 34 -> Loss: 0.153146550059 Epoch: 34 -> Test Accuracy: 77.89 Epoch: 35 -> Loss: 0.0208696424961 Epoch: 35 -> Test Accuracy: 78.92 Epoch: 36 -> Loss: 0.0984496027231 Epoch: 36 -> Test Accuracy: 79.71 Epoch: 37 -> Loss: 0.0195654183626 Epoch: 37 -> Test Accuracy: 79.92 Epoch: 38 -> Loss: 0.0948273837566 Epoch: 38 -> Test Accuracy: 79.98 Epoch: 39 -> Loss: 0.0658591911197 Epoch: 39 -> Test Accuracy: 80.08 Epoch: 40 -> Loss: 0.0876868665218 Epoch: 40 -> Test Accuracy: 80.02 Epoch: 41 -> Loss: 0.0620073378086 Epoch: 41 -> Test Accuracy: 79.98 Epoch: 42 -> Loss: 0.0212653577328 Epoch: 42 -> Test Accuracy: 80.16 Epoch: 43 -> Loss: 0.0179961770773 Epoch: 43 -> Test Accuracy: 80.03 Epoch: 44 -> Loss: 0.0110031217337 Epoch: 44 -> Test Accuracy: 79.95 Epoch: 45 -> Loss: 0.00997103750706 Epoch: 45 -> Test Accuracy: 80.04 Epoch: 46 -> Loss: 0.00899240374565 Epoch: 46 -> Test Accuracy: 80.09 Epoch: 47 -> Loss: 0.0105749368668 Epoch: 47 -> Test Accuracy: 80.0 Epoch: 48 -> Loss: 0.0594830662012 Epoch: 48 -> Test Accuracy: 80.25 Epoch: 49 -> Loss: 0.00699333846569 Epoch: 49 -> Test Accuracy: 80.16 Epoch: 50 -> Loss: 0.0160281956196 Epoch: 50 -> Test Accuracy: 80.26 Epoch: 51 -> Loss: 0.0268976092339 Epoch: 51 -> Test Accuracy: 80.28 Epoch: 52 -> Loss: 0.0339367985725 Epoch: 52 -> Test Accuracy: 80.0 Epoch: 53 -> Loss: 0.0193541646004 Epoch: 53 -> Test Accuracy: 80.12 Epoch: 54 -> Loss: 0.0446498543024 Epoch: 54 -> Test Accuracy: 80.04 Epoch: 55 -> Loss: 0.0253207534552 Epoch: 55 -> Test Accuracy: 79.98 Epoch: 56 -> Loss: 0.00760178267956 Epoch: 56 -> Test Accuracy: 80.08 Epoch: 57 -> Loss: 0.116646245122 Epoch: 57 -> Test Accuracy: 80.19 Epoch: 58 -> Loss: 0.0476978570223 Epoch: 58 -> Test Accuracy: 79.99 Epoch: 59 -> Loss: 0.0428672581911 Epoch: 59 -> Test Accuracy: 79.79 Epoch: 60 -> Loss: 0.0110652148724 Epoch: 60 -> Test Accuracy: 79.96 Epoch: 61 -> Loss: 0.0361245274544 Epoch: 61 -> Test Accuracy: 80.15 Epoch: 62 -> Loss: 0.00841024518013 Epoch: 62 -> Test Accuracy: 80.16 Epoch: 63 -> Loss: 0.0088861733675 Epoch: 63 -> Test Accuracy: 80.05 Epoch: 64 -> Loss: 0.00655518472195 Epoch: 64 -> Test Accuracy: 80.16 Epoch: 65 -> Loss: 0.0232560932636 Epoch: 65 -> Test Accuracy: 80.17 Epoch: 66 -> Loss: 0.0160289555788 Epoch: 66 -> Test Accuracy: 80.22 Epoch: 67 -> Loss: 0.0497677177191 Epoch: 67 -> Test Accuracy: 80.26 Epoch: 68 -> Loss: 0.0101454406977 Epoch: 68 -> Test Accuracy: 80.25 Epoch: 69 -> Loss: 0.0333657115698 Epoch: 69 -> Test Accuracy: 80.2 Epoch: 70 -> Loss: 0.0100089907646 Epoch: 70 -> Test Accuracy: 80.3 Epoch: 71 -> Loss: 0.0527093932033 Epoch: 71 -> Test Accuracy: 80.39 Epoch: 72 -> Loss: 0.0253549814224 Epoch: 72 -> Test Accuracy: 80.43 Epoch: 73 -> Loss: 0.0171993970871 Epoch: 73 -> Test Accuracy: 80.31 Epoch: 74 -> Loss: 0.0125417411327 Epoch: 74 -> Test Accuracy: 80.37 Epoch: 75 -> Loss: 0.0132678896189 Epoch: 75 -> Test Accuracy: 80.34 Epoch: 76 -> Loss: 0.0169373005629 Epoch: 76 -> Test Accuracy: 80.35 Epoch: 77 -> Loss: 0.0184165239334 Epoch: 77 -> Test Accuracy: 80.35 Epoch: 78 -> Loss: 0.0186169743538 Epoch: 78 -> Test Accuracy: 80.32 Epoch: 79 -> Loss: 0.0897953137755 Epoch: 79 -> Test Accuracy: 80.24 Epoch: 80 -> Loss: 0.0688715130091 Epoch: 80 -> Test Accuracy: 80.2 Epoch: 81 -> Loss: 0.0400617569685 Epoch: 81 -> Test Accuracy: 80.26 Epoch: 82 -> Loss: 0.0169174075127 Epoch: 82 -> Test Accuracy: 80.34 Epoch: 83 -> Loss: 0.0112945735455 Epoch: 83 -> Test Accuracy: 80.33 Epoch: 84 -> Loss: 0.0744367539883 Epoch: 84 -> Test Accuracy: 80.25 Epoch: 85 -> Loss: 0.0124695748091 Epoch: 85 -> Test Accuracy: 80.18 Epoch: 86 -> Loss: 0.00609792768955 Epoch: 86 -> Test Accuracy: 80.2 Epoch: 87 -> Loss: 0.0456972122192 Epoch: 87 -> Test Accuracy: 80.22 Epoch: 88 -> Loss: 0.0483660250902 Epoch: 88 -> Test Accuracy: 80.17 Epoch: 89 -> Loss: 0.0139588862658 Epoch: 89 -> Test Accuracy: 80.19 Epoch: 90 -> Loss: 0.0116586536169 Epoch: 90 -> Test Accuracy: 80.21 Epoch: 91 -> Loss: 0.00879560410976 Epoch: 91 -> Test Accuracy: 80.21 Epoch: 92 -> Loss: 0.0147249400616 Epoch: 92 -> Test Accuracy: 80.2 Epoch: 93 -> Loss: 0.0143815279007 Epoch: 93 -> Test Accuracy: 80.2 Epoch: 94 -> Loss: 0.00966900587082 Epoch: 94 -> Test Accuracy: 80.18 Epoch: 95 -> Loss: 0.0102178305387 Epoch: 95 -> Test Accuracy: 80.2 Epoch: 96 -> Loss: 0.0182191133499 Epoch: 96 -> Test Accuracy: 80.22 Epoch: 97 -> Loss: 0.073236182332 Epoch: 97 -> Test Accuracy: 80.21 Epoch: 98 -> Loss: 0.0175753235817 Epoch: 98 -> Test Accuracy: 80.21 Epoch: 99 -> Loss: 0.0189524143934 Epoch: 99 -> Test Accuracy: 80.2 Epoch: 100 -> Loss: 0.00667914748192 Epoch: 100 -> Test Accuracy: 80.19 Finished Training Epoch: 1 -> Loss: 1.8052713871 Epoch: 1 -> Test Accuracy: 32.78 Epoch: 2 -> Loss: 1.4549536705 Epoch: 2 -> Test Accuracy: 38.77 Epoch: 3 -> Loss: 1.74625742435 Epoch: 3 -> Test Accuracy: 42.13 Epoch: 4 -> Loss: 1.45812571049 Epoch: 4 -> Test Accuracy: 44.11 Epoch: 5 -> Loss: 1.23404753208 Epoch: 5 -> Test Accuracy: 46.83 Epoch: 6 -> Loss: 1.29505217075 Epoch: 6 -> Test Accuracy: 49.02 Epoch: 7 -> Loss: 0.808918893337 Epoch: 7 -> Test Accuracy: 44.8 Epoch: 8 -> Loss: 1.09893918037 Epoch: 8 -> Test Accuracy: 52.3 Epoch: 9 -> Loss: 1.08894097805 Epoch: 9 -> Test Accuracy: 53.42 Epoch: 10 -> Loss: 0.857489943504 Epoch: 10 -> Test Accuracy: 49.95 Epoch: 11 -> Loss: 1.08701896667 Epoch: 11 -> Test Accuracy: 54.2 Epoch: 12 -> Loss: 0.827267110348 Epoch: 12 -> Test Accuracy: 57.24 Epoch: 13 -> Loss: 1.14055275917 Epoch: 13 -> Test Accuracy: 55.29 Epoch: 14 -> Loss: 1.33272373676 Epoch: 14 -> Test Accuracy: 56.12 Epoch: 15 -> Loss: 0.502484440804 Epoch: 15 -> Test Accuracy: 56.8 Epoch: 16 -> Loss: 0.669791579247 Epoch: 16 -> Test Accuracy: 61.3 Epoch: 17 -> Loss: 0.920004606247 Epoch: 17 -> Test Accuracy: 56.2 Epoch: 18 -> Loss: 1.01290249825 Epoch: 18 -> Test Accuracy: 58.91 Epoch: 19 -> Loss: 0.661690473557 Epoch: 19 -> Test Accuracy: 60.98 Epoch: 20 -> Loss: 0.868436694145 Epoch: 20 -> Test Accuracy: 60.56 Epoch: 21 -> Loss: 0.431213438511 Epoch: 21 -> Test Accuracy: 61.11 Epoch: 22 -> Loss: 0.77817940712 Epoch: 22 -> Test Accuracy: 56.13 Epoch: 23 -> Loss: 0.620285212994 Epoch: 23 -> Test Accuracy: 55.99 Epoch: 24 -> Loss: 0.864762067795 Epoch: 24 -> Test Accuracy: 62.75 Epoch: 25 -> Loss: 1.06970930099 Epoch: 25 -> Test Accuracy: 59.68 Epoch: 26 -> Loss: 0.549698770046 Epoch: 26 -> Test Accuracy: 62.42 Epoch: 27 -> Loss: 0.580783247948 Epoch: 27 -> Test Accuracy: 64.4 Epoch: 28 -> Loss: 0.757976233959 Epoch: 28 -> Test Accuracy: 61.52 Epoch: 29 -> Loss: 0.73870486021 Epoch: 29 -> Test Accuracy: 61.59 Epoch: 30 -> Loss: 0.508637487888 Epoch: 30 -> Test Accuracy: 63.01 Epoch: 31 -> Loss: 0.331197977066 Epoch: 31 -> Test Accuracy: 62.15 Epoch: 32 -> Loss: 0.470873266459 Epoch: 32 -> Test Accuracy: 62.75 Epoch: 33 -> Loss: 0.374791145325 Epoch: 33 -> Test Accuracy: 64.56 Epoch: 34 -> Loss: 0.492821633816 Epoch: 34 -> Test Accuracy: 64.67 Epoch: 35 -> Loss: 0.345730870962 Epoch: 35 -> Test Accuracy: 64.3 Epoch: 36 -> Loss: 0.433670759201 Epoch: 36 -> Test Accuracy: 62.59 Epoch: 37 -> Loss: 0.317645341158 Epoch: 37 -> Test Accuracy: 61.45 Epoch: 38 -> Loss: 0.40732049942 Epoch: 38 -> Test Accuracy: 64.81 Epoch: 39 -> Loss: 0.366106927395 Epoch: 39 -> Test Accuracy: 64.42 Epoch: 40 -> Loss: 0.604485869408 Epoch: 40 -> Test Accuracy: 64.34 Epoch: 41 -> Loss: 0.489239275455 Epoch: 41 -> Test Accuracy: 61.53 Epoch: 42 -> Loss: 0.755823552608 Epoch: 42 -> Test Accuracy: 64.48 Epoch: 43 -> Loss: 0.270692288876 Epoch: 43 -> Test Accuracy: 66.21 Epoch: 44 -> Loss: 0.398351848125 Epoch: 44 -> Test Accuracy: 64.22 Epoch: 45 -> Loss: 0.491998851299 Epoch: 45 -> Test Accuracy: 63.57 Epoch: 46 -> Loss: 0.334801137447 Epoch: 46 -> Test Accuracy: 65.63 Epoch: 47 -> Loss: 0.571182549 Epoch: 47 -> Test Accuracy: 64.95 Epoch: 48 -> Loss: 0.336699008942 Epoch: 48 -> Test Accuracy: 64.46 Epoch: 49 -> Loss: 0.358170062304 Epoch: 49 -> Test Accuracy: 65.28 Epoch: 50 -> Loss: 0.250126063824 Epoch: 50 -> Test Accuracy: 65.64 Epoch: 51 -> Loss: 0.235494241118 Epoch: 51 -> Test Accuracy: 65.97 Epoch: 52 -> Loss: 0.175247132778 Epoch: 52 -> Test Accuracy: 66.37 Epoch: 53 -> Loss: 0.2731808424 Epoch: 53 -> Test Accuracy: 64.96 Epoch: 54 -> Loss: 0.316665738821 Epoch: 54 -> Test Accuracy: 62.65 Epoch: 55 -> Loss: 0.199551001191 Epoch: 55 -> Test Accuracy: 65.93 Epoch: 56 -> Loss: 0.148887112737 Epoch: 56 -> Test Accuracy: 66.13 Epoch: 57 -> Loss: 0.198421061039 Epoch: 57 -> Test Accuracy: 66.73 Epoch: 58 -> Loss: 0.449224352837 Epoch: 58 -> Test Accuracy: 66.5 Epoch: 59 -> Loss: 0.29995071888 Epoch: 59 -> Test Accuracy: 63.8 Epoch: 60 -> Loss: 0.257976233959 Epoch: 60 -> Test Accuracy: 67.68 Epoch: 61 -> Loss: 0.125610470772 Epoch: 61 -> Test Accuracy: 70.04 Epoch: 62 -> Loss: 0.0315936803818 Epoch: 62 -> Test Accuracy: 70.27 Epoch: 63 -> Loss: 0.202752485871 Epoch: 63 -> Test Accuracy: 70.31 Epoch: 64 -> Loss: 0.0247804373503 Epoch: 64 -> Test Accuracy: 70.41 Epoch: 65 -> Loss: 0.0378330647945 Epoch: 65 -> Test Accuracy: 70.53 Epoch: 66 -> Loss: 0.10579136014 Epoch: 66 -> Test Accuracy: 70.43 Epoch: 67 -> Loss: 0.0575700551271 Epoch: 67 -> Test Accuracy: 70.64 Epoch: 68 -> Loss: 0.0453321263194 Epoch: 68 -> Test Accuracy: 70.43 Epoch: 69 -> Loss: 0.0445282012224 Epoch: 69 -> Test Accuracy: 70.64 Epoch: 70 -> Loss: 0.0336372405291 Epoch: 70 -> Test Accuracy: 70.65 Epoch: 71 -> Loss: 0.0675330460072 Epoch: 71 -> Test Accuracy: 70.93 Epoch: 72 -> Loss: 0.0812125056982 Epoch: 72 -> Test Accuracy: 70.8 Epoch: 73 -> Loss: 0.0531407520175 Epoch: 73 -> Test Accuracy: 70.54 Epoch: 74 -> Loss: 0.0309773236513 Epoch: 74 -> Test Accuracy: 71.07 Epoch: 75 -> Loss: 0.0236945524812 Epoch: 75 -> Test Accuracy: 70.87 Epoch: 76 -> Loss: 0.0100391060114 Epoch: 76 -> Test Accuracy: 70.96 Epoch: 77 -> Loss: 0.0159396380186 Epoch: 77 -> Test Accuracy: 71.08 Epoch: 78 -> Loss: 0.0112279355526 Epoch: 78 -> Test Accuracy: 70.71 Epoch: 79 -> Loss: 0.0287714749575 Epoch: 79 -> Test Accuracy: 70.96 Epoch: 80 -> Loss: 0.176479279995 Epoch: 80 -> Test Accuracy: 71.1 Epoch: 81 -> Loss: 0.0553053840995 Epoch: 81 -> Test Accuracy: 70.19 Epoch: 82 -> Loss: 0.0222607702017 Epoch: 82 -> Test Accuracy: 71.05 Epoch: 83 -> Loss: 0.017648845911 Epoch: 83 -> Test Accuracy: 71.11 Epoch: 84 -> Loss: 0.00824658572674 Epoch: 84 -> Test Accuracy: 70.92 Epoch: 85 -> Loss: 0.0197389870882 Epoch: 85 -> Test Accuracy: 71.15 Epoch: 86 -> Loss: 0.0256101638079 Epoch: 86 -> Test Accuracy: 71.0 Epoch: 87 -> Loss: 0.0209289640188 Epoch: 87 -> Test Accuracy: 70.99 Epoch: 88 -> Loss: 0.0267375558615 Epoch: 88 -> Test Accuracy: 71.22 Epoch: 89 -> Loss: 0.0153915286064 Epoch: 89 -> Test Accuracy: 70.98 Epoch: 90 -> Loss: 0.0126494318247 Epoch: 90 -> Test Accuracy: 70.9 Epoch: 91 -> Loss: 0.0112864971161 Epoch: 91 -> Test Accuracy: 70.83 Epoch: 92 -> Loss: 0.00292167067528 Epoch: 92 -> Test Accuracy: 71.16 Epoch: 93 -> Loss: 0.017883643508 Epoch: 93 -> Test Accuracy: 71.19 Epoch: 94 -> Loss: 0.0337622314692 Epoch: 94 -> Test Accuracy: 71.23 Epoch: 95 -> Loss: 0.0120377838612 Epoch: 95 -> Test Accuracy: 71.1 Epoch: 96 -> Loss: 0.0127462744713 Epoch: 96 -> Test Accuracy: 71.11 Epoch: 97 -> Loss: 0.0171550512314 Epoch: 97 -> Test Accuracy: 71.11 Epoch: 98 -> Loss: 0.0599996745586 Epoch: 98 -> Test Accuracy: 70.76 Epoch: 99 -> Loss: 0.00941671431065 Epoch: 99 -> Test Accuracy: 70.31 Epoch: 100 -> Loss: 0.0182780325413 Epoch: 100 -> Test Accuracy: 70.87 Epoch: 101 -> Loss: 0.00950425863266 Epoch: 101 -> Test Accuracy: 70.9 Epoch: 102 -> Loss: 0.00967802107334 Epoch: 102 -> Test Accuracy: 70.83 Epoch: 103 -> Loss: 0.00724190473557 Epoch: 103 -> Test Accuracy: 70.64 Epoch: 104 -> Loss: 0.0324417054653 Epoch: 104 -> Test Accuracy: 70.91 Epoch: 105 -> Loss: 0.0169687718153 Epoch: 105 -> Test Accuracy: 71.13 Epoch: 106 -> Loss: 0.0314528793097 Epoch: 106 -> Test Accuracy: 71.1 Epoch: 107 -> Loss: 0.0214038491249 Epoch: 107 -> Test Accuracy: 70.72 Epoch: 108 -> Loss: 0.0333647802472 Epoch: 108 -> Test Accuracy: 70.83 Epoch: 109 -> Loss: 0.020948767662 Epoch: 109 -> Test Accuracy: 70.48 Epoch: 110 -> Loss: 0.0321883633733 Epoch: 110 -> Test Accuracy: 70.69 Epoch: 111 -> Loss: 0.0491126775742 Epoch: 111 -> Test Accuracy: 70.94 Epoch: 112 -> Loss: 0.0477923154831 Epoch: 112 -> Test Accuracy: 70.67 Epoch: 113 -> Loss: 0.0112421363592 Epoch: 113 -> Test Accuracy: 70.71 Epoch: 114 -> Loss: 0.0729033946991 Epoch: 114 -> Test Accuracy: 70.82 Epoch: 115 -> Loss: 0.0184053480625 Epoch: 115 -> Test Accuracy: 70.35 Epoch: 116 -> Loss: 0.0135750174522 Epoch: 116 -> Test Accuracy: 70.74 Epoch: 117 -> Loss: 0.0123312175274 Epoch: 117 -> Test Accuracy: 70.85 Epoch: 118 -> Loss: 0.0110506862402 Epoch: 118 -> Test Accuracy: 70.93 Epoch: 119 -> Loss: 0.0266265869141 Epoch: 119 -> Test Accuracy: 70.97 Epoch: 120 -> Loss: 0.00591425597668 Epoch: 120 -> Test Accuracy: 71.12 Epoch: 121 -> Loss: 0.0119008272886 Epoch: 121 -> Test Accuracy: 71.06 Epoch: 122 -> Loss: 0.0214089006186 Epoch: 122 -> Test Accuracy: 71.1 Epoch: 123 -> Loss: 0.050809442997 Epoch: 123 -> Test Accuracy: 70.96 Epoch: 124 -> Loss: 0.00507043302059 Epoch: 124 -> Test Accuracy: 70.89 Epoch: 125 -> Loss: 0.0147071629763 Epoch: 125 -> Test Accuracy: 70.99 Epoch: 126 -> Loss: 0.0269448906183 Epoch: 126 -> Test Accuracy: 71.06 Epoch: 127 -> Loss: 0.0107784122229 Epoch: 127 -> Test Accuracy: 71.08 Epoch: 128 -> Loss: 0.00431968271732 Epoch: 128 -> Test Accuracy: 71.14 Epoch: 129 -> Loss: 0.0145100802183 Epoch: 129 -> Test Accuracy: 71.1 Epoch: 130 -> Loss: 0.0196961015463 Epoch: 130 -> Test Accuracy: 71.08 Epoch: 131 -> Loss: 0.00568276643753 Epoch: 131 -> Test Accuracy: 71.23 Epoch: 132 -> Loss: 0.00373020768166 Epoch: 132 -> Test Accuracy: 71.15 Epoch: 133 -> Loss: 0.0101871192455 Epoch: 133 -> Test Accuracy: 71.14 Epoch: 134 -> Loss: 0.00699551403522 Epoch: 134 -> Test Accuracy: 71.19 Epoch: 135 -> Loss: 0.00743374228477 Epoch: 135 -> Test Accuracy: 71.2 Epoch: 136 -> Loss: 0.00587114691734 Epoch: 136 -> Test Accuracy: 71.08 Epoch: 137 -> Loss: 0.0149262398481 Epoch: 137 -> Test Accuracy: 71.32 Epoch: 138 -> Loss: 0.0398357957602 Epoch: 138 -> Test Accuracy: 71.19 Epoch: 139 -> Loss: 0.0252690389752 Epoch: 139 -> Test Accuracy: 71.1 Epoch: 140 -> Loss: 0.00424282252789 Epoch: 140 -> Test Accuracy: 71.22 Epoch: 141 -> Loss: 0.0226623117924 Epoch: 141 -> Test Accuracy: 71.23 Epoch: 142 -> Loss: 0.00834879279137 Epoch: 142 -> Test Accuracy: 71.3 Epoch: 143 -> Loss: 0.0448399782181 Epoch: 143 -> Test Accuracy: 71.22 Epoch: 144 -> Loss: 0.011642575264 Epoch: 144 -> Test Accuracy: 71.32 Epoch: 145 -> Loss: 0.00537167489529 Epoch: 145 -> Test Accuracy: 71.22 Epoch: 146 -> Loss: 0.0025320649147 Epoch: 146 -> Test Accuracy: 71.17 Epoch: 147 -> Loss: 0.0120529234409 Epoch: 147 -> Test Accuracy: 71.12 Epoch: 148 -> Loss: 0.0126224458218 Epoch: 148 -> Test Accuracy: 71.07 Epoch: 149 -> Loss: 0.00870415568352 Epoch: 149 -> Test Accuracy: 71.23 Epoch: 150 -> Loss: 0.00486120581627 Epoch: 150 -> Test Accuracy: 71.23 Epoch: 151 -> Loss: 0.0206541866064 Epoch: 151 -> Test Accuracy: 71.16 Epoch: 152 -> Loss: 0.0155073255301 Epoch: 152 -> Test Accuracy: 71.34 Epoch: 153 -> Loss: 0.00852030515671 Epoch: 153 -> Test Accuracy: 71.19 Epoch: 154 -> Loss: 0.0849987491965 Epoch: 154 -> Test Accuracy: 71.18 Epoch: 155 -> Loss: 0.0158245265484 Epoch: 155 -> Test Accuracy: 71.21 Epoch: 156 -> Loss: 0.0144473612309 Epoch: 156 -> Test Accuracy: 71.18 Epoch: 157 -> Loss: 0.0236957073212 Epoch: 157 -> Test Accuracy: 71.09 Epoch: 158 -> Loss: 0.00618058443069 Epoch: 158 -> Test Accuracy: 71.1 Epoch: 159 -> Loss: 0.00511218607426 Epoch: 159 -> Test Accuracy: 71.08 Epoch: 160 -> Loss: 0.00768600404263 Epoch: 160 -> Test Accuracy: 71.06 Epoch: 161 -> Loss: 0.0148626118898 Epoch: 161 -> Test Accuracy: 71.07 Epoch: 162 -> Loss: 0.00924523174763 Epoch: 162 -> Test Accuracy: 71.04 Epoch: 163 -> Loss: 0.0148717164993 Epoch: 163 -> Test Accuracy: 71.06 Epoch: 164 -> Loss: 0.0135444998741 Epoch: 164 -> Test Accuracy: 71.03 Epoch: 165 -> Loss: 0.00910976529121 Epoch: 165 -> Test Accuracy: 71.02 Epoch: 166 -> Loss: 0.0085741430521 Epoch: 166 -> Test Accuracy: 70.95 Epoch: 167 -> Loss: 0.00706927478313 Epoch: 167 -> Test Accuracy: 70.96 Epoch: 168 -> Loss: 0.00602598488331 Epoch: 168 -> Test Accuracy: 70.98 Epoch: 169 -> Loss: 0.0110482275486 Epoch: 169 -> Test Accuracy: 71.06 Epoch: 170 -> Loss: 0.0153047293425 Epoch: 170 -> Test Accuracy: 71.04 Epoch: 171 -> Loss: 0.0675692558289 Epoch: 171 -> Test Accuracy: 71.07 Epoch: 172 -> Loss: 0.039071649313 Epoch: 172 -> Test Accuracy: 71.06 Epoch: 173 -> Loss: 0.00860622525215 Epoch: 173 -> Test Accuracy: 71.03 Epoch: 174 -> Loss: 0.0049963593483 Epoch: 174 -> Test Accuracy: 71.03 Epoch: 175 -> Loss: 0.0280245244503 Epoch: 175 -> Test Accuracy: 71.03 Epoch: 176 -> Loss: 0.00963750481606 Epoch: 176 -> Test Accuracy: 71.03 Epoch: 177 -> Loss: 0.0141156464815 Epoch: 177 -> Test Accuracy: 71.04 Epoch: 178 -> Loss: 0.0196160078049 Epoch: 178 -> Test Accuracy: 71.13 Epoch: 179 -> Loss: 0.00304661691189 Epoch: 179 -> Test Accuracy: 71.12 Epoch: 180 -> Loss: 0.00694055855274 Epoch: 180 -> Test Accuracy: 71.1 Epoch: 181 -> Loss: 0.00450487434864 Epoch: 181 -> Test Accuracy: 71.15 Epoch: 182 -> Loss: 0.00474372506142 Epoch: 182 -> Test Accuracy: 71.2 Epoch: 183 -> Loss: 0.00594429671764 Epoch: 183 -> Test Accuracy: 71.14 Epoch: 184 -> Loss: 0.0180179476738 Epoch: 184 -> Test Accuracy: 71.1 Epoch: 185 -> Loss: 0.00465805828571 Epoch: 185 -> Test Accuracy: 71.11 Epoch: 186 -> Loss: 0.00408010184765 Epoch: 186 -> Test Accuracy: 71.14 Epoch: 187 -> Loss: 0.0138216912746 Epoch: 187 -> Test Accuracy: 71.11 Epoch: 188 -> Loss: 0.0191442668438 Epoch: 188 -> Test Accuracy: 71.12 Epoch: 189 -> Loss: 0.0226793438196 Epoch: 189 -> Test Accuracy: 71.1 Epoch: 190 -> Loss: 0.00964859127998 Epoch: 190 -> Test Accuracy: 71.12 Epoch: 191 -> Loss: 0.0184281468391 Epoch: 191 -> Test Accuracy: 71.12 Epoch: 192 -> Loss: 0.0143183246255 Epoch: 192 -> Test Accuracy: 71.08 Epoch: 193 -> Loss: 0.0082062035799 Epoch: 193 -> Test Accuracy: 71.11 Epoch: 194 -> Loss: 0.00283917784691 Epoch: 194 -> Test Accuracy: 71.13 Epoch: 195 -> Loss: 0.0825093537569 Epoch: 195 -> Test Accuracy: 71.03 Epoch: 196 -> Loss: 0.02463606745 Epoch: 196 -> Test Accuracy: 71.02 Epoch: 197 -> Loss: 0.0977230519056 Epoch: 197 -> Test Accuracy: 71.04 Epoch: 198 -> Loss: 0.019749417901 Epoch: 198 -> Test Accuracy: 71.13 Epoch: 199 -> Loss: 0.0182576626539 Epoch: 199 -> Test Accuracy: 71.19 Epoch: 200 -> Loss: 0.0168407261372 Epoch: 200 -> Test Accuracy: 71.12 Finished Training [1, 60] loss: 0.960 Epoch: 1 -> Loss: 1.38447475433 Epoch: 1 -> Test Accuracy: 73.28 [2, 60] loss: 0.571 Epoch: 2 -> Loss: 0.404527902603 Epoch: 2 -> Test Accuracy: 75.4 [3, 60] loss: 0.479 Epoch: 3 -> Loss: 0.560607910156 Epoch: 3 -> Test Accuracy: 78.34 [4, 60] loss: 0.426 Epoch: 4 -> Loss: 0.151405870914 Epoch: 4 -> Test Accuracy: 77.57 [5, 60] loss: 0.379 Epoch: 5 -> Loss: 0.439147114754 Epoch: 5 -> Test Accuracy: 79.42 [6, 60] loss: 0.350 Epoch: 6 -> Loss: 0.310629814863 Epoch: 6 -> Test Accuracy: 79.34 [7, 60] loss: 0.315 Epoch: 7 -> Loss: 0.591866612434 Epoch: 7 -> Test Accuracy: 79.06 [8, 60] loss: 0.304 Epoch: 8 -> Loss: 0.391358375549 Epoch: 8 -> Test Accuracy: 80.29 [9, 60] loss: 0.267 Epoch: 9 -> Loss: 0.220833972096 Epoch: 9 -> Test Accuracy: 80.21 [10, 60] loss: 0.255 Epoch: 10 -> Loss: 0.748624145985 Epoch: 10 -> Test Accuracy: 80.01 [11, 60] loss: 0.248 Epoch: 11 -> Loss: 0.29417347908 Epoch: 11 -> Test Accuracy: 77.72 [12, 60] loss: 0.229 Epoch: 12 -> Loss: 0.310448586941 Epoch: 12 -> Test Accuracy: 78.06 [13, 60] loss: 0.219 Epoch: 13 -> Loss: 0.148708105087 Epoch: 13 -> Test Accuracy: 80.37 [14, 60] loss: 0.181 Epoch: 14 -> Loss: 0.339802145958 Epoch: 14 -> Test Accuracy: 79.23 [15, 60] loss: 0.207 Epoch: 15 -> Loss: 0.260935544968 Epoch: 15 -> Test Accuracy: 81.01 [16, 60] loss: 0.191 Epoch: 16 -> Loss: 1.43244659901 Epoch: 16 -> Test Accuracy: 79.17 [17, 60] loss: 0.216 Epoch: 17 -> Loss: 0.360293626785 Epoch: 17 -> Test Accuracy: 80.25 [18, 60] loss: 0.168 Epoch: 18 -> Loss: 0.362103998661 Epoch: 18 -> Test Accuracy: 80.81 [19, 60] loss: 0.151 Epoch: 19 -> Loss: 0.0561759769917 Epoch: 19 -> Test Accuracy: 81.12 [20, 60] loss: 0.134 Epoch: 20 -> Loss: 0.0935906171799 Epoch: 20 -> Test Accuracy: 81.5 [21, 60] loss: 0.126 Epoch: 21 -> Loss: 0.171662181616 Epoch: 21 -> Test Accuracy: 81.99 [22, 60] loss: 0.133 Epoch: 22 -> Loss: 0.497239291668 Epoch: 22 -> Test Accuracy: 78.97 [23, 60] loss: 0.169 Epoch: 23 -> Loss: 0.62078088522 Epoch: 23 -> Test Accuracy: 79.45 [24, 60] loss: 0.179 Epoch: 24 -> Loss: 0.186175227165 Epoch: 24 -> Test Accuracy: 80.74 [25, 60] loss: 0.140 Epoch: 25 -> Loss: 0.055052369833 Epoch: 25 -> Test Accuracy: 80.33 [26, 60] loss: 0.111 Epoch: 26 -> Loss: 0.30346468091 Epoch: 26 -> Test Accuracy: 79.85 [27, 60] loss: 0.145 Epoch: 27 -> Loss: 0.2309871912 Epoch: 27 -> Test Accuracy: 78.44 [28, 60] loss: 0.142 Epoch: 28 -> Loss: 0.420533716679 Epoch: 28 -> Test Accuracy: 79.2 [29, 60] loss: 0.132 Epoch: 29 -> Loss: 0.0760518461466 Epoch: 29 -> Test Accuracy: 81.17 [30, 60] loss: 0.101 Epoch: 30 -> Loss: 0.219887733459 Epoch: 30 -> Test Accuracy: 79.51 [31, 60] loss: 0.114 Epoch: 31 -> Loss: 0.158827662468 Epoch: 31 -> Test Accuracy: 81.09 [32, 60] loss: 0.106 Epoch: 32 -> Loss: 0.0881378501654 Epoch: 32 -> Test Accuracy: 81.58 [33, 60] loss: 0.102 Epoch: 33 -> Loss: 0.239914447069 Epoch: 33 -> Test Accuracy: 80.56 [34, 60] loss: 0.123 Epoch: 34 -> Loss: 0.0889967381954 Epoch: 34 -> Test Accuracy: 82.02 [35, 60] loss: 0.099 Epoch: 35 -> Loss: 0.120920062065 Epoch: 35 -> Test Accuracy: 81.03 [36, 60] loss: 0.058 Epoch: 36 -> Loss: 0.0401159524918 Epoch: 36 -> Test Accuracy: 82.65 [37, 60] loss: 0.038 Epoch: 37 -> Loss: 0.100469261408 Epoch: 37 -> Test Accuracy: 83.34 [38, 60] loss: 0.033 Epoch: 38 -> Loss: 0.158164441586 Epoch: 38 -> Test Accuracy: 83.5 [39, 60] loss: 0.028 Epoch: 39 -> Loss: 0.0819121599197 Epoch: 39 -> Test Accuracy: 83.29 [40, 60] loss: 0.029 Epoch: 40 -> Loss: 0.0962285101414 Epoch: 40 -> Test Accuracy: 83.27 [41, 60] loss: 0.026 Epoch: 41 -> Loss: 0.148284494877 Epoch: 41 -> Test Accuracy: 83.42 [42, 60] loss: 0.025 Epoch: 42 -> Loss: 0.0750087499619 Epoch: 42 -> Test Accuracy: 83.54 [43, 60] loss: 0.023 Epoch: 43 -> Loss: 0.054880425334 Epoch: 43 -> Test Accuracy: 83.61 [44, 60] loss: 0.020 Epoch: 44 -> Loss: 0.0736004412174 Epoch: 44 -> Test Accuracy: 83.35 [45, 60] loss: 0.022 Epoch: 45 -> Loss: 0.0271249115467 Epoch: 45 -> Test Accuracy: 83.66 [46, 60] loss: 0.018 Epoch: 46 -> Loss: 0.0353977680206 Epoch: 46 -> Test Accuracy: 83.51 [47, 60] loss: 0.019 Epoch: 47 -> Loss: 0.227512463927 Epoch: 47 -> Test Accuracy: 83.55 [48, 60] loss: 0.021 Epoch: 48 -> Loss: 0.0304708480835 Epoch: 48 -> Test Accuracy: 83.54 [49, 60] loss: 0.018 Epoch: 49 -> Loss: 0.408481448889 Epoch: 49 -> Test Accuracy: 83.31 [50, 60] loss: 0.028 Epoch: 50 -> Loss: 0.0528238713741 Epoch: 50 -> Test Accuracy: 82.97 [51, 60] loss: 0.018 Epoch: 51 -> Loss: 0.0792226493359 Epoch: 51 -> Test Accuracy: 83.25 [52, 60] loss: 0.017 Epoch: 52 -> Loss: 0.0469386875629 Epoch: 52 -> Test Accuracy: 83.09 [53, 60] loss: 0.018 Epoch: 53 -> Loss: 0.153931587934 Epoch: 53 -> Test Accuracy: 83.12 [54, 60] loss: 0.023 Epoch: 54 -> Loss: 0.140061914921 Epoch: 54 -> Test Accuracy: 83.23 [55, 60] loss: 0.019 Epoch: 55 -> Loss: 0.193389177322 Epoch: 55 -> Test Accuracy: 83.54 [56, 60] loss: 0.019 Epoch: 56 -> Loss: 0.00955620408058 Epoch: 56 -> Test Accuracy: 83.5 [57, 60] loss: 0.015 Epoch: 57 -> Loss: 0.122540384531 Epoch: 57 -> Test Accuracy: 83.4 [58, 60] loss: 0.019 Epoch: 58 -> Loss: 0.29047998786 Epoch: 58 -> Test Accuracy: 83.24 [59, 60] loss: 0.022 Epoch: 59 -> Loss: 0.027036011219 Epoch: 59 -> Test Accuracy: 83.45 [60, 60] loss: 0.014 Epoch: 60 -> Loss: 0.0282133221626 Epoch: 60 -> Test Accuracy: 83.49 [61, 60] loss: 0.016 Epoch: 61 -> Loss: 0.0138196647167 Epoch: 61 -> Test Accuracy: 83.19 [62, 60] loss: 0.016 Epoch: 62 -> Loss: 0.0315833091736 Epoch: 62 -> Test Accuracy: 83.28 [63, 60] loss: 0.014 Epoch: 63 -> Loss: 0.0696693360806 Epoch: 63 -> Test Accuracy: 83.44 [64, 60] loss: 0.014 Epoch: 64 -> Loss: 0.140820860863 Epoch: 64 -> Test Accuracy: 83.12 [65, 60] loss: 0.017 Epoch: 65 -> Loss: 0.0753938853741 Epoch: 65 -> Test Accuracy: 83.09 [66, 60] loss: 0.017 Epoch: 66 -> Loss: 0.0540786981583 Epoch: 66 -> Test Accuracy: 83.33 [67, 60] loss: 0.015 Epoch: 67 -> Loss: 0.0456347167492 Epoch: 67 -> Test Accuracy: 83.41 [68, 60] loss: 0.014 Epoch: 68 -> Loss: 0.0100956559181 Epoch: 68 -> Test Accuracy: 82.99 [69, 60] loss: 0.012 Epoch: 69 -> Loss: 0.0364240407944 Epoch: 69 -> Test Accuracy: 83.42 [70, 60] loss: 0.013 Epoch: 70 -> Loss: 0.121539384127 Epoch: 70 -> Test Accuracy: 83.31 [71, 60] loss: 0.012 Epoch: 71 -> Loss: 0.175175726414 Epoch: 71 -> Test Accuracy: 83.38 [72, 60] loss: 0.012 Epoch: 72 -> Loss: 0.0455166697502 Epoch: 72 -> Test Accuracy: 83.26 [73, 60] loss: 0.013 Epoch: 73 -> Loss: 0.021762907505 Epoch: 73 -> Test Accuracy: 83.38 [74, 60] loss: 0.011 Epoch: 74 -> Loss: 0.0601827204227 Epoch: 74 -> Test Accuracy: 83.43 [75, 60] loss: 0.011 Epoch: 75 -> Loss: 0.0429280698299 Epoch: 75 -> Test Accuracy: 83.27 [76, 60] loss: 0.011 Epoch: 76 -> Loss: 0.0245901346207 Epoch: 76 -> Test Accuracy: 83.35 [77, 60] loss: 0.011 Epoch: 77 -> Loss: 0.009725689888 Epoch: 77 -> Test Accuracy: 83.41 [78, 60] loss: 0.012 Epoch: 78 -> Loss: 0.13119417429 Epoch: 78 -> Test Accuracy: 83.41 [79, 60] loss: 0.011 Epoch: 79 -> Loss: 0.0184655189514 Epoch: 79 -> Test Accuracy: 83.42 [80, 60] loss: 0.011 Epoch: 80 -> Loss: 0.00688621401787 Epoch: 80 -> Test Accuracy: 83.61 [81, 60] loss: 0.010 Epoch: 81 -> Loss: 0.0176421999931 Epoch: 81 -> Test Accuracy: 83.62 [82, 60] loss: 0.011 Epoch: 82 -> Loss: 0.0340698361397 Epoch: 82 -> Test Accuracy: 83.49 [83, 60] loss: 0.010 Epoch: 83 -> Loss: 0.0454367697239 Epoch: 83 -> Test Accuracy: 83.51 [84, 60] loss: 0.011 Epoch: 84 -> Loss: 0.0988063663244 Epoch: 84 -> Test Accuracy: 83.51 [85, 60] loss: 0.011 Epoch: 85 -> Loss: 0.0291695296764 Epoch: 85 -> Test Accuracy: 83.47 [86, 60] loss: 0.011 Epoch: 86 -> Loss: 0.0761667788029 Epoch: 86 -> Test Accuracy: 83.49 [87, 60] loss: 0.010 Epoch: 87 -> Loss: 0.0103091299534 Epoch: 87 -> Test Accuracy: 83.54 [88, 60] loss: 0.009 Epoch: 88 -> Loss: 0.0679731965065 Epoch: 88 -> Test Accuracy: 83.56 [89, 60] loss: 0.010 Epoch: 89 -> Loss: 0.0757903307676 Epoch: 89 -> Test Accuracy: 83.63 [90, 60] loss: 0.010 Epoch: 90 -> Loss: 0.0155855417252 Epoch: 90 -> Test Accuracy: 83.56 [91, 60] loss: 0.010 Epoch: 91 -> Loss: 0.0333587527275 Epoch: 91 -> Test Accuracy: 83.55 [92, 60] loss: 0.010 Epoch: 92 -> Loss: 0.472617059946 Epoch: 92 -> Test Accuracy: 83.62 [93, 60] loss: 0.011 Epoch: 93 -> Loss: 0.0198042690754 Epoch: 93 -> Test Accuracy: 83.69 [94, 60] loss: 0.010 Epoch: 94 -> Loss: 0.0703645050526 Epoch: 94 -> Test Accuracy: 83.66 [95, 60] loss: 0.010 Epoch: 95 -> Loss: 0.013804256916 Epoch: 95 -> Test Accuracy: 83.61 [96, 60] loss: 0.010 Epoch: 96 -> Loss: 0.191430211067 Epoch: 96 -> Test Accuracy: 83.63 [97, 60] loss: 0.010 Epoch: 97 -> Loss: 0.216685041785 Epoch: 97 -> Test Accuracy: 83.63 [98, 60] loss: 0.009 Epoch: 98 -> Loss: 0.00713688135147 Epoch: 98 -> Test Accuracy: 83.63 [99, 60] loss: 0.010 Epoch: 99 -> Loss: 0.221191763878 Epoch: 99 -> Test Accuracy: 83.63 [100, 60] loss: 0.010 Epoch: 100 -> Loss: 0.132191136479 Epoch: 100 -> Test Accuracy: 83.67 Finished Training [1, 60] loss: 1.871 Epoch: 1 -> Loss: 1.8187918663 Epoch: 1 -> Test Accuracy: 39.41 [2, 60] loss: 1.493 Epoch: 2 -> Loss: 1.33181905746 Epoch: 2 -> Test Accuracy: 45.18 [3, 60] loss: 1.339 Epoch: 3 -> Loss: 1.09351527691 Epoch: 3 -> Test Accuracy: 51.28 [4, 60] loss: 1.192 Epoch: 4 -> Loss: 1.72012042999 Epoch: 4 -> Test Accuracy: 49.4 [5, 60] loss: 1.110 Epoch: 5 -> Loss: 1.24345839024 Epoch: 5 -> Test Accuracy: 53.96 [6, 60] loss: 1.005 Epoch: 6 -> Loss: 0.790222465992 Epoch: 6 -> Test Accuracy: 54.27 [7, 60] loss: 0.943 Epoch: 7 -> Loss: 1.14126420021 Epoch: 7 -> Test Accuracy: 59.91 [8, 60] loss: 0.888 Epoch: 8 -> Loss: 1.16405045986 Epoch: 8 -> Test Accuracy: 58.87 [9, 60] loss: 0.829 Epoch: 9 -> Loss: 1.30415916443 Epoch: 9 -> Test Accuracy: 64.33 [10, 60] loss: 0.784 Epoch: 10 -> Loss: 0.877873182297 Epoch: 10 -> Test Accuracy: 64.45 [11, 60] loss: 0.733 Epoch: 11 -> Loss: 0.810325264931 Epoch: 11 -> Test Accuracy: 66.8 [12, 60] loss: 0.730 Epoch: 12 -> Loss: 1.31677258015 Epoch: 12 -> Test Accuracy: 68.8 [13, 60] loss: 0.700 Epoch: 13 -> Loss: 1.37249910831 Epoch: 13 -> Test Accuracy: 69.89 [14, 60] loss: 0.667 Epoch: 14 -> Loss: 0.609081327915 Epoch: 14 -> Test Accuracy: 68.89 [15, 60] loss: 0.630 Epoch: 15 -> Loss: 0.806365728378 Epoch: 15 -> Test Accuracy: 68.69 [16, 60] loss: 0.606 Epoch: 16 -> Loss: 0.833701610565 Epoch: 16 -> Test Accuracy: 67.11 [17, 60] loss: 0.587 Epoch: 17 -> Loss: 0.686511933804 Epoch: 17 -> Test Accuracy: 69.92 [18, 60] loss: 0.565 Epoch: 18 -> Loss: 0.609204292297 Epoch: 18 -> Test Accuracy: 67.65 [19, 60] loss: 0.590 Epoch: 19 -> Loss: 0.48525044322 Epoch: 19 -> Test Accuracy: 67.68 [20, 60] loss: 0.555 Epoch: 20 -> Loss: 0.468196600676 Epoch: 20 -> Test Accuracy: 73.53 [21, 60] loss: 0.522 Epoch: 21 -> Loss: 0.628668904305 Epoch: 21 -> Test Accuracy: 72.02 [22, 60] loss: 0.535 Epoch: 22 -> Loss: 1.28545033932 Epoch: 22 -> Test Accuracy: 67.62 [23, 60] loss: 0.537 Epoch: 23 -> Loss: 0.532577753067 Epoch: 23 -> Test Accuracy: 72.39 [24, 60] loss: 0.477 Epoch: 24 -> Loss: 0.578134298325 Epoch: 24 -> Test Accuracy: 72.71 [25, 60] loss: 0.480 Epoch: 25 -> Loss: 0.746399283409 Epoch: 25 -> Test Accuracy: 69.81 [26, 60] loss: 0.499 Epoch: 26 -> Loss: 0.732239842415 Epoch: 26 -> Test Accuracy: 70.0 [27, 60] loss: 0.489 Epoch: 27 -> Loss: 0.29195445776 Epoch: 27 -> Test Accuracy: 72.19 [28, 60] loss: 0.455 Epoch: 28 -> Loss: 0.272972077131 Epoch: 28 -> Test Accuracy: 71.96 [29, 60] loss: 0.412 Epoch: 29 -> Loss: 0.823468923569 Epoch: 29 -> Test Accuracy: 71.78 [30, 60] loss: 0.499 Epoch: 30 -> Loss: 0.551599740982 Epoch: 30 -> Test Accuracy: 71.11 [31, 60] loss: 0.453 Epoch: 31 -> Loss: 0.424916177988 Epoch: 31 -> Test Accuracy: 72.62 [32, 60] loss: 0.421 Epoch: 32 -> Loss: 0.611951828003 Epoch: 32 -> Test Accuracy: 71.06 [33, 60] loss: 0.436 Epoch: 33 -> Loss: 0.61200273037 Epoch: 33 -> Test Accuracy: 72.64 [34, 60] loss: 0.413 Epoch: 34 -> Loss: 0.710877656937 Epoch: 34 -> Test Accuracy: 72.42 [35, 60] loss: 0.407 Epoch: 35 -> Loss: 0.678718626499 Epoch: 35 -> Test Accuracy: 74.43 [36, 60] loss: 0.421 Epoch: 36 -> Loss: 0.393203854561 Epoch: 36 -> Test Accuracy: 69.38 [37, 60] loss: 0.425 Epoch: 37 -> Loss: 0.175001114607 Epoch: 37 -> Test Accuracy: 70.94 [38, 60] loss: 0.370 Epoch: 38 -> Loss: 0.646731853485 Epoch: 38 -> Test Accuracy: 73.25 [39, 60] loss: 0.401 Epoch: 39 -> Loss: 0.202637255192 Epoch: 39 -> Test Accuracy: 73.05 [40, 60] loss: 0.362 Epoch: 40 -> Loss: 0.484558403492 Epoch: 40 -> Test Accuracy: 70.61 [41, 60] loss: 0.359 Epoch: 41 -> Loss: 0.417449980974 Epoch: 41 -> Test Accuracy: 71.93 [42, 60] loss: 0.383 Epoch: 42 -> Loss: 0.453267484903 Epoch: 42 -> Test Accuracy: 73.02 [43, 60] loss: 0.335 Epoch: 43 -> Loss: 0.262421488762 Epoch: 43 -> Test Accuracy: 73.68 [44, 60] loss: 0.334 Epoch: 44 -> Loss: 0.369110047817 Epoch: 44 -> Test Accuracy: 74.82 [45, 60] loss: 0.339 Epoch: 45 -> Loss: 0.433916091919 Epoch: 45 -> Test Accuracy: 71.94 [46, 60] loss: 0.343 Epoch: 46 -> Loss: 0.266774237156 Epoch: 46 -> Test Accuracy: 74.85 [47, 60] loss: 0.332 Epoch: 47 -> Loss: 0.394922554493 Epoch: 47 -> Test Accuracy: 73.22 [48, 60] loss: 0.349 Epoch: 48 -> Loss: 0.722384214401 Epoch: 48 -> Test Accuracy: 73.23 [49, 60] loss: 0.367 Epoch: 49 -> Loss: 0.212504804134 Epoch: 49 -> Test Accuracy: 74.8 [50, 60] loss: 0.301 Epoch: 50 -> Loss: 0.564773797989 Epoch: 50 -> Test Accuracy: 74.47 [51, 60] loss: 0.332 Epoch: 51 -> Loss: 0.477853536606 Epoch: 51 -> Test Accuracy: 69.87 [52, 60] loss: 0.381 Epoch: 52 -> Loss: 0.591364145279 Epoch: 52 -> Test Accuracy: 71.55 [53, 60] loss: 0.333 Epoch: 53 -> Loss: 0.427986323833 Epoch: 53 -> Test Accuracy: 73.85 [54, 60] loss: 0.332 Epoch: 54 -> Loss: 0.29263690114 Epoch: 54 -> Test Accuracy: 74.28 [55, 60] loss: 0.291 Epoch: 55 -> Loss: 0.476581245661 Epoch: 55 -> Test Accuracy: 73.67 [56, 60] loss: 0.323 Epoch: 56 -> Loss: 0.380825281143 Epoch: 56 -> Test Accuracy: 75.12 [57, 60] loss: 0.315 Epoch: 57 -> Loss: 0.319384455681 Epoch: 57 -> Test Accuracy: 75.26 [58, 60] loss: 0.286 Epoch: 58 -> Loss: 0.25169968605 Epoch: 58 -> Test Accuracy: 75.65 [59, 60] loss: 0.302 Epoch: 59 -> Loss: 0.14394120872 Epoch: 59 -> Test Accuracy: 73.31 [60, 60] loss: 0.276 Epoch: 60 -> Loss: 0.502469241619 Epoch: 60 -> Test Accuracy: 72.05 [61, 60] loss: 0.160 Epoch: 61 -> Loss: 0.3072078228 Epoch: 61 -> Test Accuracy: 80.14 [62, 60] loss: 0.093 Epoch: 62 -> Loss: 0.150992333889 Epoch: 62 -> Test Accuracy: 80.71 [63, 60] loss: 0.071 Epoch: 63 -> Loss: 0.07261967659 Epoch: 63 -> Test Accuracy: 80.86 [64, 60] loss: 0.054 Epoch: 64 -> Loss: 0.199264407158 Epoch: 64 -> Test Accuracy: 80.73 [65, 60] loss: 0.058 Epoch: 65 -> Loss: 0.0318068563938 Epoch: 65 -> Test Accuracy: 80.51 [66, 60] loss: 0.046 Epoch: 66 -> Loss: 0.230934843421 Epoch: 66 -> Test Accuracy: 80.54 [67, 60] loss: 0.047 Epoch: 67 -> Loss: 0.0428005158901 Epoch: 67 -> Test Accuracy: 80.49 [68, 60] loss: 0.038 Epoch: 68 -> Loss: 0.119938582182 Epoch: 68 -> Test Accuracy: 80.62 [69, 60] loss: 0.036 Epoch: 69 -> Loss: 0.144359469414 Epoch: 69 -> Test Accuracy: 80.25 [70, 60] loss: 0.032 Epoch: 70 -> Loss: 0.0844859182835 Epoch: 70 -> Test Accuracy: 80.45 [71, 60] loss: 0.033 Epoch: 71 -> Loss: 0.056810721755 Epoch: 71 -> Test Accuracy: 80.65 [72, 60] loss: 0.030 Epoch: 72 -> Loss: 0.112621635199 Epoch: 72 -> Test Accuracy: 80.8 [73, 60] loss: 0.034 Epoch: 73 -> Loss: 0.213060438633 Epoch: 73 -> Test Accuracy: 81.03 [74, 60] loss: 0.033 Epoch: 74 -> Loss: 0.06031447649 Epoch: 74 -> Test Accuracy: 80.91 [75, 60] loss: 0.029 Epoch: 75 -> Loss: 0.000595450401306 Epoch: 75 -> Test Accuracy: 80.92 [76, 60] loss: 0.022 Epoch: 76 -> Loss: 0.193061798811 Epoch: 76 -> Test Accuracy: 81.13 [77, 60] loss: 0.032 Epoch: 77 -> Loss: 0.168416678905 Epoch: 77 -> Test Accuracy: 80.38 [78, 60] loss: 0.035 Epoch: 78 -> Loss: 0.0620669275522 Epoch: 78 -> Test Accuracy: 80.48 [79, 60] loss: 0.025 Epoch: 79 -> Loss: 0.155465066433 Epoch: 79 -> Test Accuracy: 80.85 [80, 60] loss: 0.027 Epoch: 80 -> Loss: 0.0427515655756 Epoch: 80 -> Test Accuracy: 80.84 [81, 60] loss: 0.019 Epoch: 81 -> Loss: 0.0293017029762 Epoch: 81 -> Test Accuracy: 81.0 [82, 60] loss: 0.017 Epoch: 82 -> Loss: 0.0407817363739 Epoch: 82 -> Test Accuracy: 80.85 [83, 60] loss: 0.018 Epoch: 83 -> Loss: 0.065326333046 Epoch: 83 -> Test Accuracy: 80.76 [84, 60] loss: 0.021 Epoch: 84 -> Loss: 0.0887884497643 Epoch: 84 -> Test Accuracy: 80.73 [85, 60] loss: 0.025 Epoch: 85 -> Loss: 0.0692778229713 Epoch: 85 -> Test Accuracy: 80.6 [86, 60] loss: 0.018 Epoch: 86 -> Loss: 0.0024850666523 Epoch: 86 -> Test Accuracy: 80.59 [87, 60] loss: 0.016 Epoch: 87 -> Loss: 0.0340867340565 Epoch: 87 -> Test Accuracy: 80.41 [88, 60] loss: 0.017 Epoch: 88 -> Loss: 0.177219048142 Epoch: 88 -> Test Accuracy: 80.26 [89, 60] loss: 0.026 Epoch: 89 -> Loss: 0.0228471755981 Epoch: 89 -> Test Accuracy: 80.79 [90, 60] loss: 0.014 Epoch: 90 -> Loss: 0.0797606706619 Epoch: 90 -> Test Accuracy: 80.63 [91, 60] loss: 0.019 Epoch: 91 -> Loss: 0.0710729658604 Epoch: 91 -> Test Accuracy: 80.91 [92, 60] loss: 0.021 Epoch: 92 -> Loss: 0.0672790110111 Epoch: 92 -> Test Accuracy: 80.81 [93, 60] loss: 0.019 Epoch: 93 -> Loss: 0.237882465124 Epoch: 93 -> Test Accuracy: 80.25 [94, 60] loss: 0.031 Epoch: 94 -> Loss: 0.0579227209091 Epoch: 94 -> Test Accuracy: 80.37 [95, 60] loss: 0.018 Epoch: 95 -> Loss: 0.158189356327 Epoch: 95 -> Test Accuracy: 80.03 [96, 60] loss: 0.029 Epoch: 96 -> Loss: 0.0466093122959 Epoch: 96 -> Test Accuracy: 80.29 [97, 60] loss: 0.017 Epoch: 97 -> Loss: 0.00992375612259 Epoch: 97 -> Test Accuracy: 80.62 [98, 60] loss: 0.013 Epoch: 98 -> Loss: 0.0287625491619 Epoch: 98 -> Test Accuracy: 80.91 [99, 60] loss: 0.014 Epoch: 99 -> Loss: 0.0344948917627 Epoch: 99 -> Test Accuracy: 80.61 [100, 60] loss: 0.015 Epoch: 100 -> Loss: 0.0827370285988 Epoch: 100 -> Test Accuracy: 80.46 [101, 60] loss: 0.023 Epoch: 101 -> Loss: 0.271556138992 Epoch: 101 -> Test Accuracy: 80.33 [102, 60] loss: 0.041 Epoch: 102 -> Loss: 0.0223966240883 Epoch: 102 -> Test Accuracy: 80.73 [103, 60] loss: 0.019 Epoch: 103 -> Loss: 0.0225029289722 Epoch: 103 -> Test Accuracy: 80.62 [104, 60] loss: 0.013 Epoch: 104 -> Loss: 0.0545526742935 Epoch: 104 -> Test Accuracy: 80.65 [105, 60] loss: 0.016 Epoch: 105 -> Loss: 0.00703835487366 Epoch: 105 -> Test Accuracy: 80.83 [106, 60] loss: 0.011 Epoch: 106 -> Loss: 0.120159447193 Epoch: 106 -> Test Accuracy: 81.06 [107, 60] loss: 0.028 Epoch: 107 -> Loss: 0.0507103055716 Epoch: 107 -> Test Accuracy: 80.4 [108, 60] loss: 0.016 Epoch: 108 -> Loss: 0.00805416703224 Epoch: 108 -> Test Accuracy: 81.16 [109, 60] loss: 0.013 Epoch: 109 -> Loss: 0.0260004997253 Epoch: 109 -> Test Accuracy: 80.19 [110, 60] loss: 0.014 Epoch: 110 -> Loss: 0.0389124155045 Epoch: 110 -> Test Accuracy: 80.75 [111, 60] loss: 0.011 Epoch: 111 -> Loss: 0.0501452684402 Epoch: 111 -> Test Accuracy: 80.69 [112, 60] loss: 0.016 Epoch: 112 -> Loss: 0.110666498542 Epoch: 112 -> Test Accuracy: 80.68 [113, 60] loss: 0.023 Epoch: 113 -> Loss: 0.278691202402 Epoch: 113 -> Test Accuracy: 79.82 [114, 60] loss: 0.042 Epoch: 114 -> Loss: 0.158509463072 Epoch: 114 -> Test Accuracy: 77.98 [115, 60] loss: 0.068 Epoch: 115 -> Loss: 0.116147324443 Epoch: 115 -> Test Accuracy: 80.06 [116, 60] loss: 0.047 Epoch: 116 -> Loss: 0.146273568273 Epoch: 116 -> Test Accuracy: 80.5 [117, 60] loss: 0.039 Epoch: 117 -> Loss: 0.098918274045 Epoch: 117 -> Test Accuracy: 79.59 [118, 60] loss: 0.031 Epoch: 118 -> Loss: 0.199494540691 Epoch: 118 -> Test Accuracy: 80.13 [119, 60] loss: 0.048 Epoch: 119 -> Loss: 0.00287753343582 Epoch: 119 -> Test Accuracy: 80.53 [120, 60] loss: 0.027 Epoch: 120 -> Loss: 0.00842189788818 Epoch: 120 -> Test Accuracy: 80.31 [121, 60] loss: 0.015 Epoch: 121 -> Loss: 0.032810986042 Epoch: 121 -> Test Accuracy: 81.07 [122, 60] loss: 0.012 Epoch: 122 -> Loss: 0.208675101399 Epoch: 122 -> Test Accuracy: 81.25 [123, 60] loss: 0.010 Epoch: 123 -> Loss: 0.0351872444153 Epoch: 123 -> Test Accuracy: 81.3 [124, 60] loss: 0.009 Epoch: 124 -> Loss: 0.0670560002327 Epoch: 124 -> Test Accuracy: 81.48 [125, 60] loss: 0.008 Epoch: 125 -> Loss: 0.0389063954353 Epoch: 125 -> Test Accuracy: 81.43 [126, 60] loss: 0.009 Epoch: 126 -> Loss: 0.0153669714928 Epoch: 126 -> Test Accuracy: 81.69 [127, 60] loss: 0.008 Epoch: 127 -> Loss: 0.281654655933 Epoch: 127 -> Test Accuracy: 81.52 [128, 60] loss: 0.009 Epoch: 128 -> Loss: 0.0480005145073 Epoch: 128 -> Test Accuracy: 81.45 [129, 60] loss: 0.009 Epoch: 129 -> Loss: 0.245687931776 Epoch: 129 -> Test Accuracy: 81.64 [130, 60] loss: 0.010 Epoch: 130 -> Loss: 0.132726043463 Epoch: 130 -> Test Accuracy: 81.44 [131, 60] loss: 0.010 Epoch: 131 -> Loss: 0.0139856338501 Epoch: 131 -> Test Accuracy: 81.35 [132, 60] loss: 0.008 Epoch: 132 -> Loss: 0.111189812422 Epoch: 132 -> Test Accuracy: 81.37 [133, 60] loss: 0.009 Epoch: 133 -> Loss: 0.0187935829163 Epoch: 133 -> Test Accuracy: 81.41 [134, 60] loss: 0.008 Epoch: 134 -> Loss: 0.0202057659626 Epoch: 134 -> Test Accuracy: 81.67 [135, 60] loss: 0.007 Epoch: 135 -> Loss: 0.343096852303 Epoch: 135 -> Test Accuracy: 81.66 [136, 60] loss: 0.013 Epoch: 136 -> Loss: 0.0472999513149 Epoch: 136 -> Test Accuracy: 81.5 [137, 60] loss: 0.008 Epoch: 137 -> Loss: 0.0110380649567 Epoch: 137 -> Test Accuracy: 81.49 [138, 60] loss: 0.007 Epoch: 138 -> Loss: 0.0304985195398 Epoch: 138 -> Test Accuracy: 81.68 [139, 60] loss: 0.007 Epoch: 139 -> Loss: 0.0988726019859 Epoch: 139 -> Test Accuracy: 81.64 [140, 60] loss: 0.008 Epoch: 140 -> Loss: 0.0773797929287 Epoch: 140 -> Test Accuracy: 81.43 [141, 60] loss: 0.008 Epoch: 141 -> Loss: 0.0299672782421 Epoch: 141 -> Test Accuracy: 81.57 [142, 60] loss: 0.007 Epoch: 142 -> Loss: 0.0310593247414 Epoch: 142 -> Test Accuracy: 81.84 [143, 60] loss: 0.006 Epoch: 143 -> Loss: 0.214785695076 Epoch: 143 -> Test Accuracy: 81.57 [144, 60] loss: 0.011 Epoch: 144 -> Loss: 0.024891063571 Epoch: 144 -> Test Accuracy: 81.44 [145, 60] loss: 0.007 Epoch: 145 -> Loss: 0.01173132658 Epoch: 145 -> Test Accuracy: 81.7 [146, 60] loss: 0.006 Epoch: 146 -> Loss: 0.0332079529762 Epoch: 146 -> Test Accuracy: 81.78 [147, 60] loss: 0.005 Epoch: 147 -> Loss: 0.00644749403 Epoch: 147 -> Test Accuracy: 81.74 [148, 60] loss: 0.005 Epoch: 148 -> Loss: 0.00720453262329 Epoch: 148 -> Test Accuracy: 81.76 [149, 60] loss: 0.006 Epoch: 149 -> Loss: 0.135471731424 Epoch: 149 -> Test Accuracy: 81.89 [150, 60] loss: 0.008 Epoch: 150 -> Loss: 0.00206413865089 Epoch: 150 -> Test Accuracy: 81.7 [151, 60] loss: 0.005 Epoch: 151 -> Loss: 0.358989804983 Epoch: 151 -> Test Accuracy: 81.44 [152, 60] loss: 0.012 Epoch: 152 -> Loss: 0.148856312037 Epoch: 152 -> Test Accuracy: 80.7 [153, 60] loss: 0.009 Epoch: 153 -> Loss: 0.00978496670723 Epoch: 153 -> Test Accuracy: 81.46 [154, 60] loss: 0.006 Epoch: 154 -> Loss: 0.027430742979 Epoch: 154 -> Test Accuracy: 81.44 [155, 60] loss: 0.006 Epoch: 155 -> Loss: 0.012020200491 Epoch: 155 -> Test Accuracy: 81.44 [156, 60] loss: 0.006 Epoch: 156 -> Loss: 0.168206393719 Epoch: 156 -> Test Accuracy: 81.46 [157, 60] loss: 0.009 Epoch: 157 -> Loss: 0.106841191649 Epoch: 157 -> Test Accuracy: 81.24 [158, 60] loss: 0.007 Epoch: 158 -> Loss: 0.0978338718414 Epoch: 158 -> Test Accuracy: 81.34 [159, 60] loss: 0.007 Epoch: 159 -> Loss: 0.0140427649021 Epoch: 159 -> Test Accuracy: 81.4 [160, 60] loss: 0.005 Epoch: 160 -> Loss: 0.141021832824 Epoch: 160 -> Test Accuracy: 81.51 [161, 60] loss: 0.006 Epoch: 161 -> Loss: 0.154865205288 Epoch: 161 -> Test Accuracy: 81.46 [162, 60] loss: 0.006 Epoch: 162 -> Loss: 0.016271084547 Epoch: 162 -> Test Accuracy: 81.43 [163, 60] loss: 0.006 Epoch: 163 -> Loss: 0.017687857151 Epoch: 163 -> Test Accuracy: 81.48 [164, 60] loss: 0.006 Epoch: 164 -> Loss: 0.0833943784237 Epoch: 164 -> Test Accuracy: 81.48 [165, 60] loss: 0.006 Epoch: 165 -> Loss: 0.112250342965 Epoch: 165 -> Test Accuracy: 81.5 [166, 60] loss: 0.005 Epoch: 166 -> Loss: 0.350206434727 Epoch: 166 -> Test Accuracy: 81.46 [167, 60] loss: 0.005 Epoch: 167 -> Loss: 0.09053170681 Epoch: 167 -> Test Accuracy: 81.35 [168, 60] loss: 0.006 Epoch: 168 -> Loss: 0.0827461481094 Epoch: 168 -> Test Accuracy: 81.41 [169, 60] loss: 0.006 Epoch: 169 -> Loss: 0.00756841897964 Epoch: 169 -> Test Accuracy: 81.34 [170, 60] loss: 0.005 Epoch: 170 -> Loss: 0.00479540228844 Epoch: 170 -> Test Accuracy: 81.45 [171, 60] loss: 0.005 Epoch: 171 -> Loss: 0.0254593491554 Epoch: 171 -> Test Accuracy: 81.39 [172, 60] loss: 0.005 Epoch: 172 -> Loss: 0.00625348091125 Epoch: 172 -> Test Accuracy: 81.46 [173, 60] loss: 0.005 Epoch: 173 -> Loss: 0.0276162922382 Epoch: 173 -> Test Accuracy: 81.48 [174, 60] loss: 0.005 Epoch: 174 -> Loss: 0.0146744251251 Epoch: 174 -> Test Accuracy: 81.59 [175, 60] loss: 0.005 Epoch: 175 -> Loss: 0.0544751882553 Epoch: 175 -> Test Accuracy: 81.67 [176, 60] loss: 0.005 Epoch: 176 -> Loss: 0.124981999397 Epoch: 176 -> Test Accuracy: 81.62 [177, 60] loss: 0.005 Epoch: 177 -> Loss: 0.00501990318298 Epoch: 177 -> Test Accuracy: 81.56 [178, 60] loss: 0.005 Epoch: 178 -> Loss: 0.227210313082 Epoch: 178 -> Test Accuracy: 81.58 [179, 60] loss: 0.006 Epoch: 179 -> Loss: 0.128590375185 Epoch: 179 -> Test Accuracy: 81.57 [180, 60] loss: 0.005 Epoch: 180 -> Loss: 0.0267832577229 Epoch: 180 -> Test Accuracy: 81.5 [181, 60] loss: 0.005 Epoch: 181 -> Loss: 0.229974016547 Epoch: 181 -> Test Accuracy: 81.63 [182, 60] loss: 0.006 Epoch: 182 -> Loss: 0.00808680057526 Epoch: 182 -> Test Accuracy: 81.52 [183, 60] loss: 0.005 Epoch: 183 -> Loss: 0.183491572738 Epoch: 183 -> Test Accuracy: 81.66 [184, 60] loss: 0.005 Epoch: 184 -> Loss: 0.0509166121483 Epoch: 184 -> Test Accuracy: 81.61 [185, 60] loss: 0.005 Epoch: 185 -> Loss: 0.00560408830643 Epoch: 185 -> Test Accuracy: 81.6 [186, 60] loss: 0.005 Epoch: 186 -> Loss: 0.0282002687454 Epoch: 186 -> Test Accuracy: 81.57 [187, 60] loss: 0.005 Epoch: 187 -> Loss: 0.210005417466 Epoch: 187 -> Test Accuracy: 81.58 [188, 60] loss: 0.005 Epoch: 188 -> Loss: 0.00581669807434 Epoch: 188 -> Test Accuracy: 81.48 [189, 60] loss: 0.005 Epoch: 189 -> Loss: 0.0183884501457 Epoch: 189 -> Test Accuracy: 81.54 [190, 60] loss: 0.005 Epoch: 190 -> Loss: 0.0202942788601 Epoch: 190 -> Test Accuracy: 81.54 [191, 60] loss: 0.005 Epoch: 191 -> Loss: 0.017327696085 Epoch: 191 -> Test Accuracy: 81.6 [192, 60] loss: 0.005 Epoch: 192 -> Loss: 0.0275225937366 Epoch: 192 -> Test Accuracy: 81.5 [193, 60] loss: 0.005 Epoch: 193 -> Loss: 0.00995624065399 Epoch: 193 -> Test Accuracy: 81.47 [194, 60] loss: 0.005 Epoch: 194 -> Loss: 0.0317635238171 Epoch: 194 -> Test Accuracy: 81.57 [195, 60] loss: 0.005 Epoch: 195 -> Loss: 0.0142703950405 Epoch: 195 -> Test Accuracy: 81.55 [196, 60] loss: 0.005 Epoch: 196 -> Loss: 0.0245959460735 Epoch: 196 -> Test Accuracy: 81.52 [197, 60] loss: 0.005 Epoch: 197 -> Loss: 0.0137966871262 Epoch: 197 -> Test Accuracy: 81.5 [198, 60] loss: 0.005 Epoch: 198 -> Loss: 0.0700696408749 Epoch: 198 -> Test Accuracy: 81.45 [199, 60] loss: 0.005 Epoch: 199 -> Loss: 0.10655759275 Epoch: 199 -> Test Accuracy: 81.4 [200, 60] loss: 0.005 Epoch: 200 -> Loss: 0.02673214674 Epoch: 200 -> Test Accuracy: 81.37 Finished Training [1, 60] loss: 0.923 [1, 120] loss: 0.626 [1, 180] loss: 0.586 [1, 240] loss: 0.536 [1, 300] loss: 0.528 [1, 360] loss: 0.496 Epoch: 1 -> Loss: 0.472029685974 Epoch: 1 -> Test Accuracy: 80.88 [2, 60] loss: 0.439 [2, 120] loss: 0.437 [2, 180] loss: 0.432 [2, 240] loss: 0.454 [2, 300] loss: 0.436 [2, 360] loss: 0.431 Epoch: 2 -> Loss: 0.282150655985 Epoch: 2 -> Test Accuracy: 82.76 [3, 60] loss: 0.375 [3, 120] loss: 0.395 [3, 180] loss: 0.386 [3, 240] loss: 0.411 [3, 300] loss: 0.370 [3, 360] loss: 0.391 Epoch: 3 -> Loss: 0.309844821692 Epoch: 3 -> Test Accuracy: 83.21 [4, 60] loss: 0.351 [4, 120] loss: 0.366 [4, 180] loss: 0.361 [4, 240] loss: 0.362 [4, 300] loss: 0.377 [4, 360] loss: 0.361 Epoch: 4 -> Loss: 0.358151316643 Epoch: 4 -> Test Accuracy: 84.23 [5, 60] loss: 0.331 [5, 120] loss: 0.331 [5, 180] loss: 0.352 [5, 240] loss: 0.343 [5, 300] loss: 0.339 [5, 360] loss: 0.359 Epoch: 5 -> Loss: 0.207544609904 Epoch: 5 -> Test Accuracy: 84.41 [6, 60] loss: 0.313 [6, 120] loss: 0.338 [6, 180] loss: 0.322 [6, 240] loss: 0.331 [6, 300] loss: 0.340 [6, 360] loss: 0.333 Epoch: 6 -> Loss: 0.278534770012 Epoch: 6 -> Test Accuracy: 84.73 [7, 60] loss: 0.295 [7, 120] loss: 0.318 [7, 180] loss: 0.335 [7, 240] loss: 0.317 [7, 300] loss: 0.321 [7, 360] loss: 0.340 Epoch: 7 -> Loss: 0.221092373133 Epoch: 7 -> Test Accuracy: 85.06 [8, 60] loss: 0.289 [8, 120] loss: 0.299 [8, 180] loss: 0.311 [8, 240] loss: 0.316 [8, 300] loss: 0.321 [8, 360] loss: 0.336 Epoch: 8 -> Loss: 0.341584444046 Epoch: 8 -> Test Accuracy: 84.65 [9, 60] loss: 0.282 [9, 120] loss: 0.281 [9, 180] loss: 0.301 [9, 240] loss: 0.342 [9, 300] loss: 0.314 [9, 360] loss: 0.322 Epoch: 9 -> Loss: 0.425087928772 Epoch: 9 -> Test Accuracy: 85.42 [10, 60] loss: 0.277 [10, 120] loss: 0.284 [10, 180] loss: 0.283 [10, 240] loss: 0.295 [10, 300] loss: 0.329 [10, 360] loss: 0.305 Epoch: 10 -> Loss: 0.128976032138 Epoch: 10 -> Test Accuracy: 85.01 [11, 60] loss: 0.275 [11, 120] loss: 0.270 [11, 180] loss: 0.301 [11, 240] loss: 0.298 [11, 300] loss: 0.308 [11, 360] loss: 0.314 Epoch: 11 -> Loss: 0.211427256465 Epoch: 11 -> Test Accuracy: 84.63 [12, 60] loss: 0.279 [12, 120] loss: 0.266 [12, 180] loss: 0.303 [12, 240] loss: 0.291 [12, 300] loss: 0.303 [12, 360] loss: 0.288 Epoch: 12 -> Loss: 0.353486508131 Epoch: 12 -> Test Accuracy: 85.5 [13, 60] loss: 0.264 [13, 120] loss: 0.272 [13, 180] loss: 0.293 [13, 240] loss: 0.308 [13, 300] loss: 0.294 [13, 360] loss: 0.297 Epoch: 13 -> Loss: 0.226709887385 Epoch: 13 -> Test Accuracy: 85.62 [14, 60] loss: 0.263 [14, 120] loss: 0.269 [14, 180] loss: 0.278 [14, 240] loss: 0.290 [14, 300] loss: 0.290 [14, 360] loss: 0.303 Epoch: 14 -> Loss: 0.284619033337 Epoch: 14 -> Test Accuracy: 85.12 [15, 60] loss: 0.273 [15, 120] loss: 0.288 [15, 180] loss: 0.279 [15, 240] loss: 0.285 [15, 300] loss: 0.292 [15, 360] loss: 0.280 Epoch: 15 -> Loss: 0.357054203749 Epoch: 15 -> Test Accuracy: 85.63 [16, 60] loss: 0.256 [16, 120] loss: 0.281 [16, 180] loss: 0.278 [16, 240] loss: 0.279 [16, 300] loss: 0.295 [16, 360] loss: 0.296 Epoch: 16 -> Loss: 0.431780010462 Epoch: 16 -> Test Accuracy: 85.53 [17, 60] loss: 0.257 [17, 120] loss: 0.264 [17, 180] loss: 0.286 [17, 240] loss: 0.285 [17, 300] loss: 0.277 [17, 360] loss: 0.297 Epoch: 17 -> Loss: 0.428378582001 Epoch: 17 -> Test Accuracy: 86.07 [18, 60] loss: 0.259 [18, 120] loss: 0.262 [18, 180] loss: 0.281 [18, 240] loss: 0.272 [18, 300] loss: 0.287 [18, 360] loss: 0.277 Epoch: 18 -> Loss: 0.195079699159 Epoch: 18 -> Test Accuracy: 85.85 [19, 60] loss: 0.256 [19, 120] loss: 0.275 [19, 180] loss: 0.267 [19, 240] loss: 0.279 [19, 300] loss: 0.283 [19, 360] loss: 0.278 Epoch: 19 -> Loss: 0.194859102368 Epoch: 19 -> Test Accuracy: 85.32 [20, 60] loss: 0.238 [20, 120] loss: 0.260 [20, 180] loss: 0.285 [20, 240] loss: 0.281 [20, 300] loss: 0.282 [20, 360] loss: 0.276 Epoch: 20 -> Loss: 0.205233901739 Epoch: 20 -> Test Accuracy: 86.19 [21, 60] loss: 0.238 [21, 120] loss: 0.263 [21, 180] loss: 0.282 [21, 240] loss: 0.288 [21, 300] loss: 0.271 [21, 360] loss: 0.298 Epoch: 21 -> Loss: 0.30584281683 Epoch: 21 -> Test Accuracy: 85.64 [22, 60] loss: 0.258 [22, 120] loss: 0.263 [22, 180] loss: 0.259 [22, 240] loss: 0.269 [22, 300] loss: 0.278 [22, 360] loss: 0.288 Epoch: 22 -> Loss: 0.258524358273 Epoch: 22 -> Test Accuracy: 85.85 [23, 60] loss: 0.254 [23, 120] loss: 0.254 [23, 180] loss: 0.276 [23, 240] loss: 0.284 [23, 300] loss: 0.281 [23, 360] loss: 0.268 Epoch: 23 -> Loss: 0.32976347208 Epoch: 23 -> Test Accuracy: 85.31 [24, 60] loss: 0.257 [24, 120] loss: 0.244 [24, 180] loss: 0.278 [24, 240] loss: 0.276 [24, 300] loss: 0.272 [24, 360] loss: 0.266 Epoch: 24 -> Loss: 0.312492758036 Epoch: 24 -> Test Accuracy: 86.34 [25, 60] loss: 0.254 [25, 120] loss: 0.250 [25, 180] loss: 0.257 [25, 240] loss: 0.266 [25, 300] loss: 0.283 [25, 360] loss: 0.285 Epoch: 25 -> Loss: 0.302136838436 Epoch: 25 -> Test Accuracy: 86.08 [26, 60] loss: 0.248 [26, 120] loss: 0.250 [26, 180] loss: 0.259 [26, 240] loss: 0.258 [26, 300] loss: 0.290 [26, 360] loss: 0.270 Epoch: 26 -> Loss: 0.273910939693 Epoch: 26 -> Test Accuracy: 85.72 [27, 60] loss: 0.234 [27, 120] loss: 0.248 [27, 180] loss: 0.265 [27, 240] loss: 0.278 [27, 300] loss: 0.265 [27, 360] loss: 0.287 Epoch: 27 -> Loss: 0.265882998705 Epoch: 27 -> Test Accuracy: 86.02 [28, 60] loss: 0.236 [28, 120] loss: 0.235 [28, 180] loss: 0.263 [28, 240] loss: 0.254 [28, 300] loss: 0.263 [28, 360] loss: 0.286 Epoch: 28 -> Loss: 0.269533962011 Epoch: 28 -> Test Accuracy: 85.89 [29, 60] loss: 0.246 [29, 120] loss: 0.257 [29, 180] loss: 0.264 [29, 240] loss: 0.258 [29, 300] loss: 0.266 [29, 360] loss: 0.272 Epoch: 29 -> Loss: 0.421383947134 Epoch: 29 -> Test Accuracy: 85.39 [30, 60] loss: 0.245 [30, 120] loss: 0.251 [30, 180] loss: 0.265 [30, 240] loss: 0.265 [30, 300] loss: 0.281 [30, 360] loss: 0.278 Epoch: 30 -> Loss: 0.290724217892 Epoch: 30 -> Test Accuracy: 85.6 [31, 60] loss: 0.246 [31, 120] loss: 0.255 [31, 180] loss: 0.271 [31, 240] loss: 0.260 [31, 300] loss: 0.265 [31, 360] loss: 0.270 Epoch: 31 -> Loss: 0.24599532783 Epoch: 31 -> Test Accuracy: 86.12 [32, 60] loss: 0.240 [32, 120] loss: 0.257 [32, 180] loss: 0.272 [32, 240] loss: 0.270 [32, 300] loss: 0.250 [32, 360] loss: 0.270 Epoch: 32 -> Loss: 0.198343589902 Epoch: 32 -> Test Accuracy: 86.15 [33, 60] loss: 0.240 [33, 120] loss: 0.254 [33, 180] loss: 0.253 [33, 240] loss: 0.269 [33, 300] loss: 0.279 [33, 360] loss: 0.267 Epoch: 33 -> Loss: 0.337949573994 Epoch: 33 -> Test Accuracy: 85.44 [34, 60] loss: 0.235 [34, 120] loss: 0.241 [34, 180] loss: 0.261 [34, 240] loss: 0.283 [34, 300] loss: 0.272 [34, 360] loss: 0.270 Epoch: 34 -> Loss: 0.290144622326 Epoch: 34 -> Test Accuracy: 86.09 [35, 60] loss: 0.243 [35, 120] loss: 0.235 [35, 180] loss: 0.250 [35, 240] loss: 0.264 [35, 300] loss: 0.283 [35, 360] loss: 0.278 Epoch: 35 -> Loss: 0.235561653972 Epoch: 35 -> Test Accuracy: 86.24 [36, 60] loss: 0.214 [36, 120] loss: 0.166 [36, 180] loss: 0.180 [36, 240] loss: 0.162 [36, 300] loss: 0.161 [36, 360] loss: 0.168 Epoch: 36 -> Loss: 0.140135973692 Epoch: 36 -> Test Accuracy: 88.48 [37, 60] loss: 0.150 [37, 120] loss: 0.148 [37, 180] loss: 0.146 [37, 240] loss: 0.141 [37, 300] loss: 0.141 [37, 360] loss: 0.142 Epoch: 37 -> Loss: 0.127807617188 Epoch: 37 -> Test Accuracy: 88.56 [38, 60] loss: 0.123 [38, 120] loss: 0.130 [38, 180] loss: 0.129 [38, 240] loss: 0.142 [38, 300] loss: 0.135 [38, 360] loss: 0.136 Epoch: 38 -> Loss: 0.172238379717 Epoch: 38 -> Test Accuracy: 88.61 [39, 60] loss: 0.115 [39, 120] loss: 0.118 [39, 180] loss: 0.123 [39, 240] loss: 0.131 [39, 300] loss: 0.131 [39, 360] loss: 0.127 Epoch: 39 -> Loss: 0.13024738431 Epoch: 39 -> Test Accuracy: 88.1 [40, 60] loss: 0.111 [40, 120] loss: 0.110 [40, 180] loss: 0.118 [40, 240] loss: 0.126 [40, 300] loss: 0.123 [40, 360] loss: 0.124 Epoch: 40 -> Loss: 0.0988387987018 Epoch: 40 -> Test Accuracy: 88.52 [41, 60] loss: 0.107 [41, 120] loss: 0.105 [41, 180] loss: 0.115 [41, 240] loss: 0.114 [41, 300] loss: 0.117 [41, 360] loss: 0.126 Epoch: 41 -> Loss: 0.127695798874 Epoch: 41 -> Test Accuracy: 88.43 [42, 60] loss: 0.104 [42, 120] loss: 0.108 [42, 180] loss: 0.103 [42, 240] loss: 0.115 [42, 300] loss: 0.123 [42, 360] loss: 0.124 Epoch: 42 -> Loss: 0.114689387381 Epoch: 42 -> Test Accuracy: 88.14 [43, 60] loss: 0.100 [43, 120] loss: 0.096 [43, 180] loss: 0.104 [43, 240] loss: 0.112 [43, 300] loss: 0.117 [43, 360] loss: 0.118 Epoch: 43 -> Loss: 0.160857588053 Epoch: 43 -> Test Accuracy: 87.7 [44, 60] loss: 0.101 [44, 120] loss: 0.096 [44, 180] loss: 0.106 [44, 240] loss: 0.118 [44, 300] loss: 0.112 [44, 360] loss: 0.113 Epoch: 44 -> Loss: 0.0992850586772 Epoch: 44 -> Test Accuracy: 87.96 [45, 60] loss: 0.090 [45, 120] loss: 0.097 [45, 180] loss: 0.106 [45, 240] loss: 0.107 [45, 300] loss: 0.110 [45, 360] loss: 0.112 Epoch: 45 -> Loss: 0.119655810297 Epoch: 45 -> Test Accuracy: 87.45 [46, 60] loss: 0.097 [46, 120] loss: 0.104 [46, 180] loss: 0.100 [46, 240] loss: 0.118 [46, 300] loss: 0.112 [46, 360] loss: 0.111 Epoch: 46 -> Loss: 0.101536586881 Epoch: 46 -> Test Accuracy: 88.22 [47, 60] loss: 0.090 [47, 120] loss: 0.107 [47, 180] loss: 0.102 [47, 240] loss: 0.106 [47, 300] loss: 0.113 [47, 360] loss: 0.116 Epoch: 47 -> Loss: 0.100572682917 Epoch: 47 -> Test Accuracy: 87.81 [48, 60] loss: 0.104 [48, 120] loss: 0.097 [48, 180] loss: 0.103 [48, 240] loss: 0.100 [48, 300] loss: 0.115 [48, 360] loss: 0.113 Epoch: 48 -> Loss: 0.132347792387 Epoch: 48 -> Test Accuracy: 87.6 [49, 60] loss: 0.096 [49, 120] loss: 0.101 [49, 180] loss: 0.103 [49, 240] loss: 0.113 [49, 300] loss: 0.117 [49, 360] loss: 0.118 Epoch: 49 -> Loss: 0.0958237200975 Epoch: 49 -> Test Accuracy: 88.02 [50, 60] loss: 0.105 [50, 120] loss: 0.101 [50, 180] loss: 0.101 [50, 240] loss: 0.111 [50, 300] loss: 0.107 [50, 360] loss: 0.111 Epoch: 50 -> Loss: 0.12903265655 Epoch: 50 -> Test Accuracy: 87.72 [51, 60] loss: 0.097 [51, 120] loss: 0.106 [51, 180] loss: 0.099 [51, 240] loss: 0.111 [51, 300] loss: 0.100 [51, 360] loss: 0.115 Epoch: 51 -> Loss: 0.153112858534 Epoch: 51 -> Test Accuracy: 87.64 [52, 60] loss: 0.102 [52, 120] loss: 0.103 [52, 180] loss: 0.100 [52, 240] loss: 0.110 [52, 300] loss: 0.114 [52, 360] loss: 0.112 Epoch: 52 -> Loss: 0.112087152898 Epoch: 52 -> Test Accuracy: 87.61 [53, 60] loss: 0.098 [53, 120] loss: 0.099 [53, 180] loss: 0.106 [53, 240] loss: 0.106 [53, 300] loss: 0.118 [53, 360] loss: 0.116 Epoch: 53 -> Loss: 0.17863933742 Epoch: 53 -> Test Accuracy: 87.71 [54, 60] loss: 0.091 [54, 120] loss: 0.107 [54, 180] loss: 0.103 [54, 240] loss: 0.109 [54, 300] loss: 0.109 [54, 360] loss: 0.109 Epoch: 54 -> Loss: 0.0908921808004 Epoch: 54 -> Test Accuracy: 87.58 [55, 60] loss: 0.104 [55, 120] loss: 0.096 [55, 180] loss: 0.103 [55, 240] loss: 0.108 [55, 300] loss: 0.109 [55, 360] loss: 0.116 Epoch: 55 -> Loss: 0.0838604420424 Epoch: 55 -> Test Accuracy: 87.76 [56, 60] loss: 0.092 [56, 120] loss: 0.103 [56, 180] loss: 0.100 [56, 240] loss: 0.110 [56, 300] loss: 0.107 [56, 360] loss: 0.118 Epoch: 56 -> Loss: 0.196120411158 Epoch: 56 -> Test Accuracy: 86.96 [57, 60] loss: 0.103 [57, 120] loss: 0.103 [57, 180] loss: 0.097 [57, 240] loss: 0.109 [57, 300] loss: 0.112 [57, 360] loss: 0.121 Epoch: 57 -> Loss: 0.110311388969 Epoch: 57 -> Test Accuracy: 87.64 [58, 60] loss: 0.095 [58, 120] loss: 0.103 [58, 180] loss: 0.104 [58, 240] loss: 0.109 [58, 300] loss: 0.119 [58, 360] loss: 0.110 Epoch: 58 -> Loss: 0.113412238657 Epoch: 58 -> Test Accuracy: 87.1 [59, 60] loss: 0.108 [59, 120] loss: 0.106 [59, 180] loss: 0.099 [59, 240] loss: 0.106 [59, 300] loss: 0.110 [59, 360] loss: 0.106 Epoch: 59 -> Loss: 0.136997550726 Epoch: 59 -> Test Accuracy: 87.46 [60, 60] loss: 0.098 [60, 120] loss: 0.100 [60, 180] loss: 0.108 [60, 240] loss: 0.103 [60, 300] loss: 0.107 [60, 360] loss: 0.111 Epoch: 60 -> Loss: 0.0945501327515 Epoch: 60 -> Test Accuracy: 87.62 [61, 60] loss: 0.099 [61, 120] loss: 0.099 [61, 180] loss: 0.108 [61, 240] loss: 0.105 [61, 300] loss: 0.116 [61, 360] loss: 0.112 Epoch: 61 -> Loss: 0.141589984298 Epoch: 61 -> Test Accuracy: 86.68 [62, 60] loss: 0.108 [62, 120] loss: 0.100 [62, 180] loss: 0.096 [62, 240] loss: 0.105 [62, 300] loss: 0.106 [62, 360] loss: 0.112 Epoch: 62 -> Loss: 0.12697981298 Epoch: 62 -> Test Accuracy: 87.42 [63, 60] loss: 0.093 [63, 120] loss: 0.095 [63, 180] loss: 0.103 [63, 240] loss: 0.107 [63, 300] loss: 0.105 [63, 360] loss: 0.118 Epoch: 63 -> Loss: 0.105083033442 Epoch: 63 -> Test Accuracy: 87.24 [64, 60] loss: 0.090 [64, 120] loss: 0.096 [64, 180] loss: 0.102 [64, 240] loss: 0.102 [64, 300] loss: 0.105 [64, 360] loss: 0.109 Epoch: 64 -> Loss: 0.0736970603466 Epoch: 64 -> Test Accuracy: 87.56 [65, 60] loss: 0.095 [65, 120] loss: 0.095 [65, 180] loss: 0.094 [65, 240] loss: 0.104 [65, 300] loss: 0.111 [65, 360] loss: 0.108 Epoch: 65 -> Loss: 0.125747174025 Epoch: 65 -> Test Accuracy: 87.38 [66, 60] loss: 0.099 [66, 120] loss: 0.099 [66, 180] loss: 0.097 [66, 240] loss: 0.113 [66, 300] loss: 0.108 [66, 360] loss: 0.115 Epoch: 66 -> Loss: 0.145292937756 Epoch: 66 -> Test Accuracy: 87.94 [67, 60] loss: 0.097 [67, 120] loss: 0.097 [67, 180] loss: 0.101 [67, 240] loss: 0.100 [67, 300] loss: 0.107 [67, 360] loss: 0.109 Epoch: 67 -> Loss: 0.208483934402 Epoch: 67 -> Test Accuracy: 87.41 [68, 60] loss: 0.100 [68, 120] loss: 0.102 [68, 180] loss: 0.103 [68, 240] loss: 0.101 [68, 300] loss: 0.112 [68, 360] loss: 0.105 Epoch: 68 -> Loss: 0.197883352637 Epoch: 68 -> Test Accuracy: 87.42 [69, 60] loss: 0.099 [69, 120] loss: 0.100 [69, 180] loss: 0.097 [69, 240] loss: 0.104 [69, 300] loss: 0.106 [69, 360] loss: 0.115 Epoch: 69 -> Loss: 0.0952011197805 Epoch: 69 -> Test Accuracy: 86.91 [70, 60] loss: 0.094 [70, 120] loss: 0.104 [70, 180] loss: 0.100 [70, 240] loss: 0.104 [70, 300] loss: 0.113 [70, 360] loss: 0.110 Epoch: 70 -> Loss: 0.148934438825 Epoch: 70 -> Test Accuracy: 87.7 [71, 60] loss: 0.078 [71, 120] loss: 0.071 [71, 180] loss: 0.067 [71, 240] loss: 0.061 [71, 300] loss: 0.059 [71, 360] loss: 0.064 Epoch: 71 -> Loss: 0.101838968694 Epoch: 71 -> Test Accuracy: 88.89 [72, 60] loss: 0.054 [72, 120] loss: 0.050 [72, 180] loss: 0.055 [72, 240] loss: 0.055 [72, 300] loss: 0.053 [72, 360] loss: 0.055 Epoch: 72 -> Loss: 0.0629889145494 Epoch: 72 -> Test Accuracy: 88.74 [73, 60] loss: 0.048 [73, 120] loss: 0.048 [73, 180] loss: 0.049 [73, 240] loss: 0.046 [73, 300] loss: 0.050 [73, 360] loss: 0.046 Epoch: 73 -> Loss: 0.0920459479094 Epoch: 73 -> Test Accuracy: 88.85 [74, 60] loss: 0.043 [74, 120] loss: 0.045 [74, 180] loss: 0.045 [74, 240] loss: 0.044 [74, 300] loss: 0.049 [74, 360] loss: 0.042 Epoch: 74 -> Loss: 0.0641450062394 Epoch: 74 -> Test Accuracy: 89.06 [75, 60] loss: 0.040 [75, 120] loss: 0.041 [75, 180] loss: 0.046 [75, 240] loss: 0.043 [75, 300] loss: 0.046 [75, 360] loss: 0.038 Epoch: 75 -> Loss: 0.0862347185612 Epoch: 75 -> Test Accuracy: 88.87 [76, 60] loss: 0.041 [76, 120] loss: 0.040 [76, 180] loss: 0.041 [76, 240] loss: 0.040 [76, 300] loss: 0.043 [76, 360] loss: 0.042 Epoch: 76 -> Loss: 0.031208762899 Epoch: 76 -> Test Accuracy: 88.99 [77, 60] loss: 0.037 [77, 120] loss: 0.036 [77, 180] loss: 0.036 [77, 240] loss: 0.037 [77, 300] loss: 0.039 [77, 360] loss: 0.039 Epoch: 77 -> Loss: 0.0608403757215 Epoch: 77 -> Test Accuracy: 88.89 [78, 60] loss: 0.036 [78, 120] loss: 0.036 [78, 180] loss: 0.038 [78, 240] loss: 0.037 [78, 300] loss: 0.034 [78, 360] loss: 0.041 Epoch: 78 -> Loss: 0.0351943150163 Epoch: 78 -> Test Accuracy: 88.95 [79, 60] loss: 0.034 [79, 120] loss: 0.034 [79, 180] loss: 0.038 [79, 240] loss: 0.040 [79, 300] loss: 0.037 [79, 360] loss: 0.037 Epoch: 79 -> Loss: 0.0699258968234 Epoch: 79 -> Test Accuracy: 89.22 [80, 60] loss: 0.037 [80, 120] loss: 0.032 [80, 180] loss: 0.031 [80, 240] loss: 0.032 [80, 300] loss: 0.035 [80, 360] loss: 0.033 Epoch: 80 -> Loss: 0.0586454160511 Epoch: 80 -> Test Accuracy: 89.04 [81, 60] loss: 0.033 [81, 120] loss: 0.034 [81, 180] loss: 0.034 [81, 240] loss: 0.033 [81, 300] loss: 0.034 [81, 360] loss: 0.033 Epoch: 81 -> Loss: 0.0648641735315 Epoch: 81 -> Test Accuracy: 88.74 [82, 60] loss: 0.030 [82, 120] loss: 0.036 [82, 180] loss: 0.033 [82, 240] loss: 0.035 [82, 300] loss: 0.032 [82, 360] loss: 0.036 Epoch: 82 -> Loss: 0.0335661396384 Epoch: 82 -> Test Accuracy: 88.82 [83, 60] loss: 0.031 [83, 120] loss: 0.030 [83, 180] loss: 0.035 [83, 240] loss: 0.034 [83, 300] loss: 0.035 [83, 360] loss: 0.034 Epoch: 83 -> Loss: 0.0633686929941 Epoch: 83 -> Test Accuracy: 88.84 [84, 60] loss: 0.031 [84, 120] loss: 0.029 [84, 180] loss: 0.031 [84, 240] loss: 0.032 [84, 300] loss: 0.034 [84, 360] loss: 0.032 Epoch: 84 -> Loss: 0.0487174540758 Epoch: 84 -> Test Accuracy: 88.78 [85, 60] loss: 0.030 [85, 120] loss: 0.029 [85, 180] loss: 0.032 [85, 240] loss: 0.030 [85, 300] loss: 0.031 [85, 360] loss: 0.031 Epoch: 85 -> Loss: 0.0835116282105 Epoch: 85 -> Test Accuracy: 88.71 [86, 60] loss: 0.028 [86, 120] loss: 0.029 [86, 180] loss: 0.031 [86, 240] loss: 0.029 [86, 300] loss: 0.027 [86, 360] loss: 0.028 Epoch: 86 -> Loss: 0.023648628965 Epoch: 86 -> Test Accuracy: 88.91 [87, 60] loss: 0.029 [87, 120] loss: 0.031 [87, 180] loss: 0.029 [87, 240] loss: 0.027 [87, 300] loss: 0.026 [87, 360] loss: 0.028 Epoch: 87 -> Loss: 0.0327759198844 Epoch: 87 -> Test Accuracy: 89.0 [88, 60] loss: 0.028 [88, 120] loss: 0.028 [88, 180] loss: 0.026 [88, 240] loss: 0.027 [88, 300] loss: 0.027 [88, 360] loss: 0.028 Epoch: 88 -> Loss: 0.0510449707508 Epoch: 88 -> Test Accuracy: 88.97 [89, 60] loss: 0.026 [89, 120] loss: 0.027 [89, 180] loss: 0.028 [89, 240] loss: 0.025 [89, 300] loss: 0.028 [89, 360] loss: 0.027 Epoch: 89 -> Loss: 0.0274531003088 Epoch: 89 -> Test Accuracy: 89.13 [90, 60] loss: 0.026 [90, 120] loss: 0.028 [90, 180] loss: 0.025 [90, 240] loss: 0.026 [90, 300] loss: 0.026 [90, 360] loss: 0.027 Epoch: 90 -> Loss: 0.0168426446617 Epoch: 90 -> Test Accuracy: 88.93 [91, 60] loss: 0.028 [91, 120] loss: 0.025 [91, 180] loss: 0.026 [91, 240] loss: 0.025 [91, 300] loss: 0.027 [91, 360] loss: 0.029 Epoch: 91 -> Loss: 0.0277911536396 Epoch: 91 -> Test Accuracy: 88.97 [92, 60] loss: 0.029 [92, 120] loss: 0.026 [92, 180] loss: 0.025 [92, 240] loss: 0.026 [92, 300] loss: 0.028 [92, 360] loss: 0.027 Epoch: 92 -> Loss: 0.02224926278 Epoch: 92 -> Test Accuracy: 88.85 [93, 60] loss: 0.025 [93, 120] loss: 0.026 [93, 180] loss: 0.027 [93, 240] loss: 0.025 [93, 300] loss: 0.026 [93, 360] loss: 0.027 Epoch: 93 -> Loss: 0.0203959103674 Epoch: 93 -> Test Accuracy: 88.94 [94, 60] loss: 0.026 [94, 120] loss: 0.027 [94, 180] loss: 0.026 [94, 240] loss: 0.025 [94, 300] loss: 0.025 [94, 360] loss: 0.026 Epoch: 94 -> Loss: 0.0144302491099 Epoch: 94 -> Test Accuracy: 88.98 [95, 60] loss: 0.024 [95, 120] loss: 0.025 [95, 180] loss: 0.027 [95, 240] loss: 0.027 [95, 300] loss: 0.025 [95, 360] loss: 0.025 Epoch: 95 -> Loss: 0.0329773649573 Epoch: 95 -> Test Accuracy: 89.0 [96, 60] loss: 0.026 [96, 120] loss: 0.027 [96, 180] loss: 0.025 [96, 240] loss: 0.024 [96, 300] loss: 0.024 [96, 360] loss: 0.026 Epoch: 96 -> Loss: 0.0315780416131 Epoch: 96 -> Test Accuracy: 89.0 [97, 60] loss: 0.027 [97, 120] loss: 0.027 [97, 180] loss: 0.025 [97, 240] loss: 0.026 [97, 300] loss: 0.026 [97, 360] loss: 0.023 Epoch: 97 -> Loss: 0.0222754776478 Epoch: 97 -> Test Accuracy: 89.04 [98, 60] loss: 0.026 [98, 120] loss: 0.026 [98, 180] loss: 0.025 [98, 240] loss: 0.024 [98, 300] loss: 0.026 [98, 360] loss: 0.028 Epoch: 98 -> Loss: 0.0145429130644 Epoch: 98 -> Test Accuracy: 89.01 [99, 60] loss: 0.024 [99, 120] loss: 0.024 [99, 180] loss: 0.026 [99, 240] loss: 0.023 [99, 300] loss: 0.026 [99, 360] loss: 0.026 Epoch: 99 -> Loss: 0.0245034340769 Epoch: 99 -> Test Accuracy: 88.99 [100, 60] loss: 0.023 [100, 120] loss: 0.024 [100, 180] loss: 0.026 [100, 240] loss: 0.025 [100, 300] loss: 0.023 [100, 360] loss: 0.027 Epoch: 100 -> Loss: 0.01750581339 Epoch: 100 -> Test Accuracy: 88.94 Finished Training [1, 60] loss: 1.810 [1, 120] loss: 1.521 [1, 180] loss: 1.352 [1, 240] loss: 1.236 [1, 300] loss: 1.153 [1, 360] loss: 1.105 Epoch: 1 -> Loss: 1.12288761139 Epoch: 1 -> Test Accuracy: 61.0 [2, 60] loss: 1.007 [2, 120] loss: 0.947 [2, 180] loss: 0.907 [2, 240] loss: 0.896 [2, 300] loss: 0.846 [2, 360] loss: 0.813 Epoch: 2 -> Loss: 0.656613349915 Epoch: 2 -> Test Accuracy: 70.11 [3, 60] loss: 0.786 [3, 120] loss: 0.770 [3, 180] loss: 0.763 [3, 240] loss: 0.736 [3, 300] loss: 0.731 [3, 360] loss: 0.722 Epoch: 3 -> Loss: 0.771119475365 Epoch: 3 -> Test Accuracy: 72.99 [4, 60] loss: 0.691 [4, 120] loss: 0.696 [4, 180] loss: 0.661 [4, 240] loss: 0.679 [4, 300] loss: 0.662 [4, 360] loss: 0.667 Epoch: 4 -> Loss: 0.700662791729 Epoch: 4 -> Test Accuracy: 73.77 [5, 60] loss: 0.613 [5, 120] loss: 0.643 [5, 180] loss: 0.630 [5, 240] loss: 0.646 [5, 300] loss: 0.628 [5, 360] loss: 0.606 Epoch: 5 -> Loss: 0.581497311592 Epoch: 5 -> Test Accuracy: 76.96 [6, 60] loss: 0.593 [6, 120] loss: 0.573 [6, 180] loss: 0.614 [6, 240] loss: 0.606 [6, 300] loss: 0.609 [6, 360] loss: 0.583 Epoch: 6 -> Loss: 0.478937298059 Epoch: 6 -> Test Accuracy: 78.15 [7, 60] loss: 0.559 [7, 120] loss: 0.586 [7, 180] loss: 0.561 [7, 240] loss: 0.565 [7, 300] loss: 0.570 [7, 360] loss: 0.561 Epoch: 7 -> Loss: 0.330141574144 Epoch: 7 -> Test Accuracy: 78.88 [8, 60] loss: 0.549 [8, 120] loss: 0.544 [8, 180] loss: 0.570 [8, 240] loss: 0.536 [8, 300] loss: 0.545 [8, 360] loss: 0.535 Epoch: 8 -> Loss: 0.493169873953 Epoch: 8 -> Test Accuracy: 78.33 [9, 60] loss: 0.514 [9, 120] loss: 0.529 [9, 180] loss: 0.538 [9, 240] loss: 0.531 [9, 300] loss: 0.518 [9, 360] loss: 0.549 Epoch: 9 -> Loss: 0.515562534332 Epoch: 9 -> Test Accuracy: 80.23 [10, 60] loss: 0.504 [10, 120] loss: 0.511 [10, 180] loss: 0.525 [10, 240] loss: 0.517 [10, 300] loss: 0.501 [10, 360] loss: 0.547 Epoch: 10 -> Loss: 0.511408030987 Epoch: 10 -> Test Accuracy: 81.49 [11, 60] loss: 0.479 [11, 120] loss: 0.500 [11, 180] loss: 0.528 [11, 240] loss: 0.524 [11, 300] loss: 0.506 [11, 360] loss: 0.511 Epoch: 11 -> Loss: 0.641905009747 Epoch: 11 -> Test Accuracy: 80.95 [12, 60] loss: 0.474 [12, 120] loss: 0.513 [12, 180] loss: 0.489 [12, 240] loss: 0.488 [12, 300] loss: 0.496 [12, 360] loss: 0.492 Epoch: 12 -> Loss: 0.483081430197 Epoch: 12 -> Test Accuracy: 79.52 [13, 60] loss: 0.486 [13, 120] loss: 0.490 [13, 180] loss: 0.477 [13, 240] loss: 0.492 [13, 300] loss: 0.500 [13, 360] loss: 0.509 Epoch: 13 -> Loss: 0.418179035187 Epoch: 13 -> Test Accuracy: 80.94 [14, 60] loss: 0.480 [14, 120] loss: 0.457 [14, 180] loss: 0.470 [14, 240] loss: 0.491 [14, 300] loss: 0.487 [14, 360] loss: 0.496 Epoch: 14 -> Loss: 0.465601056814 Epoch: 14 -> Test Accuracy: 81.38 [15, 60] loss: 0.458 [15, 120] loss: 0.475 [15, 180] loss: 0.459 [15, 240] loss: 0.468 [15, 300] loss: 0.476 [15, 360] loss: 0.470 Epoch: 15 -> Loss: 0.502071738243 Epoch: 15 -> Test Accuracy: 80.78 [16, 60] loss: 0.454 [16, 120] loss: 0.455 [16, 180] loss: 0.452 [16, 240] loss: 0.466 [16, 300] loss: 0.470 [16, 360] loss: 0.483 Epoch: 16 -> Loss: 0.539138913155 Epoch: 16 -> Test Accuracy: 83.08 [17, 60] loss: 0.433 [17, 120] loss: 0.466 [17, 180] loss: 0.457 [17, 240] loss: 0.467 [17, 300] loss: 0.456 [17, 360] loss: 0.457 Epoch: 17 -> Loss: 0.438570439816 Epoch: 17 -> Test Accuracy: 82.02 [18, 60] loss: 0.445 [18, 120] loss: 0.459 [18, 180] loss: 0.449 [18, 240] loss: 0.457 [18, 300] loss: 0.474 [18, 360] loss: 0.450 Epoch: 18 -> Loss: 0.288309037685 Epoch: 18 -> Test Accuracy: 81.83 [19, 60] loss: 0.423 [19, 120] loss: 0.457 [19, 180] loss: 0.447 [19, 240] loss: 0.436 [19, 300] loss: 0.455 [19, 360] loss: 0.465 Epoch: 19 -> Loss: 0.568620026112 Epoch: 19 -> Test Accuracy: 81.61 [20, 60] loss: 0.426 [20, 120] loss: 0.443 [20, 180] loss: 0.424 [20, 240] loss: 0.455 [20, 300] loss: 0.458 [20, 360] loss: 0.475 Epoch: 20 -> Loss: 0.423507988453 Epoch: 20 -> Test Accuracy: 82.19 [21, 60] loss: 0.415 [21, 120] loss: 0.432 [21, 180] loss: 0.428 [21, 240] loss: 0.451 [21, 300] loss: 0.453 [21, 360] loss: 0.452 Epoch: 21 -> Loss: 0.330270588398 Epoch: 21 -> Test Accuracy: 83.3 [22, 60] loss: 0.413 [22, 120] loss: 0.413 [22, 180] loss: 0.437 [22, 240] loss: 0.440 [22, 300] loss: 0.432 [22, 360] loss: 0.473 Epoch: 22 -> Loss: 0.324644684792 Epoch: 22 -> Test Accuracy: 83.41 [23, 60] loss: 0.407 [23, 120] loss: 0.437 [23, 180] loss: 0.453 [23, 240] loss: 0.418 [23, 300] loss: 0.440 [23, 360] loss: 0.422 Epoch: 23 -> Loss: 0.340217113495 Epoch: 23 -> Test Accuracy: 82.2 [24, 60] loss: 0.406 [24, 120] loss: 0.439 [24, 180] loss: 0.414 [24, 240] loss: 0.446 [24, 300] loss: 0.440 [24, 360] loss: 0.427 Epoch: 24 -> Loss: 0.404260098934 Epoch: 24 -> Test Accuracy: 82.32 [25, 60] loss: 0.407 [25, 120] loss: 0.424 [25, 180] loss: 0.417 [25, 240] loss: 0.428 [25, 300] loss: 0.420 [25, 360] loss: 0.465 Epoch: 25 -> Loss: 0.548399567604 Epoch: 25 -> Test Accuracy: 81.93 [26, 60] loss: 0.420 [26, 120] loss: 0.395 [26, 180] loss: 0.443 [26, 240] loss: 0.441 [26, 300] loss: 0.440 [26, 360] loss: 0.426 Epoch: 26 -> Loss: 0.495992511511 Epoch: 26 -> Test Accuracy: 81.87 [27, 60] loss: 0.405 [27, 120] loss: 0.429 [27, 180] loss: 0.414 [27, 240] loss: 0.419 [27, 300] loss: 0.452 [27, 360] loss: 0.446 Epoch: 27 -> Loss: 0.423348963261 Epoch: 27 -> Test Accuracy: 83.15 [28, 60] loss: 0.399 [28, 120] loss: 0.402 [28, 180] loss: 0.411 [28, 240] loss: 0.418 [28, 300] loss: 0.431 [28, 360] loss: 0.423 Epoch: 28 -> Loss: 0.42018944025 Epoch: 28 -> Test Accuracy: 83.09 [29, 60] loss: 0.382 [29, 120] loss: 0.425 [29, 180] loss: 0.398 [29, 240] loss: 0.444 [29, 300] loss: 0.424 [29, 360] loss: 0.435 Epoch: 29 -> Loss: 0.463393777609 Epoch: 29 -> Test Accuracy: 81.38 [30, 60] loss: 0.392 [30, 120] loss: 0.412 [30, 180] loss: 0.428 [30, 240] loss: 0.402 [30, 300] loss: 0.423 [30, 360] loss: 0.426 Epoch: 30 -> Loss: 0.498779624701 Epoch: 30 -> Test Accuracy: 82.5 [31, 60] loss: 0.384 [31, 120] loss: 0.398 [31, 180] loss: 0.408 [31, 240] loss: 0.406 [31, 300] loss: 0.422 [31, 360] loss: 0.426 Epoch: 31 -> Loss: 0.430401235819 Epoch: 31 -> Test Accuracy: 82.81 [32, 60] loss: 0.402 [32, 120] loss: 0.413 [32, 180] loss: 0.431 [32, 240] loss: 0.403 [32, 300] loss: 0.411 [32, 360] loss: 0.447 Epoch: 32 -> Loss: 0.481689304113 Epoch: 32 -> Test Accuracy: 83.93 [33, 60] loss: 0.389 [33, 120] loss: 0.419 [33, 180] loss: 0.420 [33, 240] loss: 0.423 [33, 300] loss: 0.413 [33, 360] loss: 0.410 Epoch: 33 -> Loss: 0.509442150593 Epoch: 33 -> Test Accuracy: 82.36 [34, 60] loss: 0.388 [34, 120] loss: 0.387 [34, 180] loss: 0.401 [34, 240] loss: 0.435 [34, 300] loss: 0.441 [34, 360] loss: 0.410 Epoch: 34 -> Loss: 0.367752313614 Epoch: 34 -> Test Accuracy: 83.79 [35, 60] loss: 0.395 [35, 120] loss: 0.405 [35, 180] loss: 0.400 [35, 240] loss: 0.407 [35, 300] loss: 0.398 [35, 360] loss: 0.426 Epoch: 35 -> Loss: 0.374817669392 Epoch: 35 -> Test Accuracy: 83.0 [36, 60] loss: 0.388 [36, 120] loss: 0.399 [36, 180] loss: 0.414 [36, 240] loss: 0.405 [36, 300] loss: 0.434 [36, 360] loss: 0.414 Epoch: 36 -> Loss: 0.409789174795 Epoch: 36 -> Test Accuracy: 84.48 [37, 60] loss: 0.363 [37, 120] loss: 0.391 [37, 180] loss: 0.417 [37, 240] loss: 0.404 [37, 300] loss: 0.431 [37, 360] loss: 0.423 Epoch: 37 -> Loss: 0.465282261372 Epoch: 37 -> Test Accuracy: 82.01 [38, 60] loss: 0.355 [38, 120] loss: 0.406 [38, 180] loss: 0.405 [38, 240] loss: 0.400 [38, 300] loss: 0.407 [38, 360] loss: 0.420 Epoch: 38 -> Loss: 0.604222893715 Epoch: 38 -> Test Accuracy: 82.33 [39, 60] loss: 0.385 [39, 120] loss: 0.390 [39, 180] loss: 0.407 [39, 240] loss: 0.414 [39, 300] loss: 0.409 [39, 360] loss: 0.391 Epoch: 39 -> Loss: 0.261452525854 Epoch: 39 -> Test Accuracy: 83.29 [40, 60] loss: 0.395 [40, 120] loss: 0.381 [40, 180] loss: 0.397 [40, 240] loss: 0.400 [40, 300] loss: 0.411 [40, 360] loss: 0.408 Epoch: 40 -> Loss: 0.410374075174 Epoch: 40 -> Test Accuracy: 80.89 [41, 60] loss: 0.382 [41, 120] loss: 0.387 [41, 180] loss: 0.401 [41, 240] loss: 0.424 [41, 300] loss: 0.399 [41, 360] loss: 0.386 Epoch: 41 -> Loss: 0.302482843399 Epoch: 41 -> Test Accuracy: 84.02 [42, 60] loss: 0.375 [42, 120] loss: 0.383 [42, 180] loss: 0.402 [42, 240] loss: 0.414 [42, 300] loss: 0.410 [42, 360] loss: 0.401 Epoch: 42 -> Loss: 0.545179009438 Epoch: 42 -> Test Accuracy: 83.73 [43, 60] loss: 0.395 [43, 120] loss: 0.394 [43, 180] loss: 0.405 [43, 240] loss: 0.398 [43, 300] loss: 0.403 [43, 360] loss: 0.404 Epoch: 43 -> Loss: 0.430734246969 Epoch: 43 -> Test Accuracy: 83.44 [44, 60] loss: 0.387 [44, 120] loss: 0.382 [44, 180] loss: 0.404 [44, 240] loss: 0.406 [44, 300] loss: 0.398 [44, 360] loss: 0.399 Epoch: 44 -> Loss: 0.45417919755 Epoch: 44 -> Test Accuracy: 83.89 [45, 60] loss: 0.383 [45, 120] loss: 0.388 [45, 180] loss: 0.388 [45, 240] loss: 0.388 [45, 300] loss: 0.391 [45, 360] loss: 0.427 Epoch: 45 -> Loss: 0.483777672052 Epoch: 45 -> Test Accuracy: 83.69 [46, 60] loss: 0.352 [46, 120] loss: 0.418 [46, 180] loss: 0.401 [46, 240] loss: 0.402 [46, 300] loss: 0.406 [46, 360] loss: 0.397 Epoch: 46 -> Loss: 0.354863882065 Epoch: 46 -> Test Accuracy: 84.82 [47, 60] loss: 0.363 [47, 120] loss: 0.398 [47, 180] loss: 0.382 [47, 240] loss: 0.391 [47, 300] loss: 0.427 [47, 360] loss: 0.410 Epoch: 47 -> Loss: 0.375812649727 Epoch: 47 -> Test Accuracy: 82.92 [48, 60] loss: 0.392 [48, 120] loss: 0.382 [48, 180] loss: 0.381 [48, 240] loss: 0.404 [48, 300] loss: 0.388 [48, 360] loss: 0.397 Epoch: 48 -> Loss: 0.652660787106 Epoch: 48 -> Test Accuracy: 83.91 [49, 60] loss: 0.382 [49, 120] loss: 0.385 [49, 180] loss: 0.371 [49, 240] loss: 0.401 [49, 300] loss: 0.417 [49, 360] loss: 0.412 Epoch: 49 -> Loss: 0.257808864117 Epoch: 49 -> Test Accuracy: 82.13 [50, 60] loss: 0.394 [50, 120] loss: 0.372 [50, 180] loss: 0.386 [50, 240] loss: 0.391 [50, 300] loss: 0.401 [50, 360] loss: 0.419 Epoch: 50 -> Loss: 0.4174708426 Epoch: 50 -> Test Accuracy: 83.12 [51, 60] loss: 0.362 [51, 120] loss: 0.365 [51, 180] loss: 0.397 [51, 240] loss: 0.401 [51, 300] loss: 0.419 [51, 360] loss: 0.415 Epoch: 51 -> Loss: 0.376387149096 Epoch: 51 -> Test Accuracy: 83.89 [52, 60] loss: 0.373 [52, 120] loss: 0.382 [52, 180] loss: 0.377 [52, 240] loss: 0.387 [52, 300] loss: 0.410 [52, 360] loss: 0.402 Epoch: 52 -> Loss: 0.488904893398 Epoch: 52 -> Test Accuracy: 84.17 [53, 60] loss: 0.369 [53, 120] loss: 0.386 [53, 180] loss: 0.395 [53, 240] loss: 0.396 [53, 300] loss: 0.393 [53, 360] loss: 0.401 Epoch: 53 -> Loss: 0.561681389809 Epoch: 53 -> Test Accuracy: 84.35 [54, 60] loss: 0.378 [54, 120] loss: 0.364 [54, 180] loss: 0.379 [54, 240] loss: 0.399 [54, 300] loss: 0.406 [54, 360] loss: 0.398 Epoch: 54 -> Loss: 0.377320796251 Epoch: 54 -> Test Accuracy: 83.94 [55, 60] loss: 0.352 [55, 120] loss: 0.368 [55, 180] loss: 0.396 [55, 240] loss: 0.395 [55, 300] loss: 0.396 [55, 360] loss: 0.403 Epoch: 55 -> Loss: 0.412592083216 Epoch: 55 -> Test Accuracy: 82.49 [56, 60] loss: 0.387 [56, 120] loss: 0.381 [56, 180] loss: 0.370 [56, 240] loss: 0.395 [56, 300] loss: 0.391 [56, 360] loss: 0.405 Epoch: 56 -> Loss: 0.258317321539 Epoch: 56 -> Test Accuracy: 83.6 [57, 60] loss: 0.376 [57, 120] loss: 0.376 [57, 180] loss: 0.408 [57, 240] loss: 0.391 [57, 300] loss: 0.386 [57, 360] loss: 0.380 Epoch: 57 -> Loss: 0.580971181393 Epoch: 57 -> Test Accuracy: 83.28 [58, 60] loss: 0.368 [58, 120] loss: 0.369 [58, 180] loss: 0.395 [58, 240] loss: 0.386 [58, 300] loss: 0.411 [58, 360] loss: 0.394 Epoch: 58 -> Loss: 0.397532403469 Epoch: 58 -> Test Accuracy: 85.16 [59, 60] loss: 0.382 [59, 120] loss: 0.384 [59, 180] loss: 0.387 [59, 240] loss: 0.392 [59, 300] loss: 0.398 [59, 360] loss: 0.390 Epoch: 59 -> Loss: 0.393878996372 Epoch: 59 -> Test Accuracy: 83.21 [60, 60] loss: 0.351 [60, 120] loss: 0.388 [60, 180] loss: 0.399 [60, 240] loss: 0.365 [60, 300] loss: 0.380 [60, 360] loss: 0.395 Epoch: 60 -> Loss: 0.27115380764 Epoch: 60 -> Test Accuracy: 84.06 [61, 60] loss: 0.271 [61, 120] loss: 0.225 [61, 180] loss: 0.204 [61, 240] loss: 0.215 [61, 300] loss: 0.201 [61, 360] loss: 0.217 Epoch: 61 -> Loss: 0.184683158994 Epoch: 61 -> Test Accuracy: 89.15 [62, 60] loss: 0.172 [62, 120] loss: 0.175 [62, 180] loss: 0.182 [62, 240] loss: 0.175 [62, 300] loss: 0.184 [62, 360] loss: 0.178 Epoch: 62 -> Loss: 0.247966215014 Epoch: 62 -> Test Accuracy: 89.64 [63, 60] loss: 0.156 [63, 120] loss: 0.147 [63, 180] loss: 0.153 [63, 240] loss: 0.175 [63, 300] loss: 0.172 [63, 360] loss: 0.171 Epoch: 63 -> Loss: 0.0811947211623 Epoch: 63 -> Test Accuracy: 89.98 [64, 60] loss: 0.147 [64, 120] loss: 0.145 [64, 180] loss: 0.159 [64, 240] loss: 0.154 [64, 300] loss: 0.152 [64, 360] loss: 0.158 Epoch: 64 -> Loss: 0.163954615593 Epoch: 64 -> Test Accuracy: 89.41 [65, 60] loss: 0.139 [65, 120] loss: 0.133 [65, 180] loss: 0.145 [65, 240] loss: 0.143 [65, 300] loss: 0.140 [65, 360] loss: 0.143 Epoch: 65 -> Loss: 0.124986365438 Epoch: 65 -> Test Accuracy: 88.97 [66, 60] loss: 0.129 [66, 120] loss: 0.129 [66, 180] loss: 0.133 [66, 240] loss: 0.142 [66, 300] loss: 0.143 [66, 360] loss: 0.143 Epoch: 66 -> Loss: 0.185143619776 Epoch: 66 -> Test Accuracy: 88.99 [67, 60] loss: 0.122 [67, 120] loss: 0.117 [67, 180] loss: 0.123 [67, 240] loss: 0.140 [67, 300] loss: 0.135 [67, 360] loss: 0.134 Epoch: 67 -> Loss: 0.167147248983 Epoch: 67 -> Test Accuracy: 88.84 [68, 60] loss: 0.121 [68, 120] loss: 0.123 [68, 180] loss: 0.155 [68, 240] loss: 0.147 [68, 300] loss: 0.140 [68, 360] loss: 0.146 Epoch: 68 -> Loss: 0.0710888057947 Epoch: 68 -> Test Accuracy: 89.42 [69, 60] loss: 0.118 [69, 120] loss: 0.127 [69, 180] loss: 0.126 [69, 240] loss: 0.134 [69, 300] loss: 0.139 [69, 360] loss: 0.142 Epoch: 69 -> Loss: 0.240266010165 Epoch: 69 -> Test Accuracy: 89.55 [70, 60] loss: 0.129 [70, 120] loss: 0.126 [70, 180] loss: 0.134 [70, 240] loss: 0.134 [70, 300] loss: 0.148 [70, 360] loss: 0.149 Epoch: 70 -> Loss: 0.102905832231 Epoch: 70 -> Test Accuracy: 88.61 [71, 60] loss: 0.128 [71, 120] loss: 0.133 [71, 180] loss: 0.125 [71, 240] loss: 0.132 [71, 300] loss: 0.140 [71, 360] loss: 0.145 Epoch: 71 -> Loss: 0.230306059122 Epoch: 71 -> Test Accuracy: 88.57 [72, 60] loss: 0.121 [72, 120] loss: 0.127 [72, 180] loss: 0.130 [72, 240] loss: 0.141 [72, 300] loss: 0.144 [72, 360] loss: 0.155 Epoch: 72 -> Loss: 0.207720682025 Epoch: 72 -> Test Accuracy: 88.41 [73, 60] loss: 0.142 [73, 120] loss: 0.131 [73, 180] loss: 0.137 [73, 240] loss: 0.133 [73, 300] loss: 0.136 [73, 360] loss: 0.149 Epoch: 73 -> Loss: 0.142775982618 Epoch: 73 -> Test Accuracy: 87.91 [74, 60] loss: 0.126 [74, 120] loss: 0.121 [74, 180] loss: 0.130 [74, 240] loss: 0.148 [74, 300] loss: 0.136 [74, 360] loss: 0.143 Epoch: 74 -> Loss: 0.116629101336 Epoch: 74 -> Test Accuracy: 88.44 [75, 60] loss: 0.127 [75, 120] loss: 0.136 [75, 180] loss: 0.136 [75, 240] loss: 0.148 [75, 300] loss: 0.148 [75, 360] loss: 0.149 Epoch: 75 -> Loss: 0.0989060997963 Epoch: 75 -> Test Accuracy: 87.99 [76, 60] loss: 0.121 [76, 120] loss: 0.126 [76, 180] loss: 0.124 [76, 240] loss: 0.135 [76, 300] loss: 0.149 [76, 360] loss: 0.152 Epoch: 76 -> Loss: 0.109110489488 Epoch: 76 -> Test Accuracy: 87.75 [77, 60] loss: 0.135 [77, 120] loss: 0.132 [77, 180] loss: 0.126 [77, 240] loss: 0.140 [77, 300] loss: 0.142 [77, 360] loss: 0.146 Epoch: 77 -> Loss: 0.111236132681 Epoch: 77 -> Test Accuracy: 88.7 [78, 60] loss: 0.130 [78, 120] loss: 0.136 [78, 180] loss: 0.130 [78, 240] loss: 0.146 [78, 300] loss: 0.141 [78, 360] loss: 0.147 Epoch: 78 -> Loss: 0.200542613864 Epoch: 78 -> Test Accuracy: 88.6 [79, 60] loss: 0.139 [79, 120] loss: 0.129 [79, 180] loss: 0.138 [79, 240] loss: 0.141 [79, 300] loss: 0.148 [79, 360] loss: 0.146 Epoch: 79 -> Loss: 0.094291254878 Epoch: 79 -> Test Accuracy: 88.41 [80, 60] loss: 0.129 [80, 120] loss: 0.130 [80, 180] loss: 0.134 [80, 240] loss: 0.139 [80, 300] loss: 0.148 [80, 360] loss: 0.146 Epoch: 80 -> Loss: 0.173632472754 Epoch: 80 -> Test Accuracy: 88.76 [81, 60] loss: 0.126 [81, 120] loss: 0.128 [81, 180] loss: 0.137 [81, 240] loss: 0.136 [81, 300] loss: 0.144 [81, 360] loss: 0.152 Epoch: 81 -> Loss: 0.279099166393 Epoch: 81 -> Test Accuracy: 89.16 [82, 60] loss: 0.138 [82, 120] loss: 0.142 [82, 180] loss: 0.152 [82, 240] loss: 0.130 [82, 300] loss: 0.152 [82, 360] loss: 0.151 Epoch: 82 -> Loss: 0.135829001665 Epoch: 82 -> Test Accuracy: 88.82 [83, 60] loss: 0.123 [83, 120] loss: 0.119 [83, 180] loss: 0.134 [83, 240] loss: 0.147 [83, 300] loss: 0.146 [83, 360] loss: 0.149 Epoch: 83 -> Loss: 0.142142474651 Epoch: 83 -> Test Accuracy: 88.2 [84, 60] loss: 0.112 [84, 120] loss: 0.137 [84, 180] loss: 0.128 [84, 240] loss: 0.138 [84, 300] loss: 0.141 [84, 360] loss: 0.153 Epoch: 84 -> Loss: 0.196907579899 Epoch: 84 -> Test Accuracy: 88.01 [85, 60] loss: 0.135 [85, 120] loss: 0.132 [85, 180] loss: 0.118 [85, 240] loss: 0.126 [85, 300] loss: 0.142 [85, 360] loss: 0.141 Epoch: 85 -> Loss: 0.189886763692 Epoch: 85 -> Test Accuracy: 87.82 [86, 60] loss: 0.125 [86, 120] loss: 0.133 [86, 180] loss: 0.132 [86, 240] loss: 0.132 [86, 300] loss: 0.149 [86, 360] loss: 0.150 Epoch: 86 -> Loss: 0.148542538285 Epoch: 86 -> Test Accuracy: 89.02 [87, 60] loss: 0.127 [87, 120] loss: 0.126 [87, 180] loss: 0.139 [87, 240] loss: 0.131 [87, 300] loss: 0.143 [87, 360] loss: 0.146 Epoch: 87 -> Loss: 0.172487735748 Epoch: 87 -> Test Accuracy: 88.54 [88, 60] loss: 0.124 [88, 120] loss: 0.120 [88, 180] loss: 0.147 [88, 240] loss: 0.134 [88, 300] loss: 0.137 [88, 360] loss: 0.132 Epoch: 88 -> Loss: 0.147699654102 Epoch: 88 -> Test Accuracy: 88.09 [89, 60] loss: 0.125 [89, 120] loss: 0.127 [89, 180] loss: 0.139 [89, 240] loss: 0.138 [89, 300] loss: 0.144 [89, 360] loss: 0.151 Epoch: 89 -> Loss: 0.229477882385 Epoch: 89 -> Test Accuracy: 87.65 [90, 60] loss: 0.136 [90, 120] loss: 0.128 [90, 180] loss: 0.129 [90, 240] loss: 0.131 [90, 300] loss: 0.139 [90, 360] loss: 0.141 Epoch: 90 -> Loss: 0.19520072639 Epoch: 90 -> Test Accuracy: 87.86 [91, 60] loss: 0.124 [91, 120] loss: 0.127 [91, 180] loss: 0.127 [91, 240] loss: 0.134 [91, 300] loss: 0.149 [91, 360] loss: 0.148 Epoch: 91 -> Loss: 0.2024397403 Epoch: 91 -> Test Accuracy: 88.52 [92, 60] loss: 0.120 [92, 120] loss: 0.122 [92, 180] loss: 0.122 [92, 240] loss: 0.127 [92, 300] loss: 0.140 [92, 360] loss: 0.151 Epoch: 92 -> Loss: 0.0761212706566 Epoch: 92 -> Test Accuracy: 88.65 [93, 60] loss: 0.118 [93, 120] loss: 0.113 [93, 180] loss: 0.133 [93, 240] loss: 0.131 [93, 300] loss: 0.141 [93, 360] loss: 0.150 Epoch: 93 -> Loss: 0.145360976458 Epoch: 93 -> Test Accuracy: 88.7 [94, 60] loss: 0.112 [94, 120] loss: 0.119 [94, 180] loss: 0.124 [94, 240] loss: 0.142 [94, 300] loss: 0.141 [94, 360] loss: 0.133 Epoch: 94 -> Loss: 0.166016787291 Epoch: 94 -> Test Accuracy: 88.45 [95, 60] loss: 0.114 [95, 120] loss: 0.124 [95, 180] loss: 0.118 [95, 240] loss: 0.137 [95, 300] loss: 0.141 [95, 360] loss: 0.146 Epoch: 95 -> Loss: 0.102204702795 Epoch: 95 -> Test Accuracy: 88.71 [96, 60] loss: 0.122 [96, 120] loss: 0.123 [96, 180] loss: 0.129 [96, 240] loss: 0.122 [96, 300] loss: 0.133 [96, 360] loss: 0.133 Epoch: 96 -> Loss: 0.108994938433 Epoch: 96 -> Test Accuracy: 88.34 [97, 60] loss: 0.122 [97, 120] loss: 0.121 [97, 180] loss: 0.125 [97, 240] loss: 0.113 [97, 300] loss: 0.139 [97, 360] loss: 0.137 Epoch: 97 -> Loss: 0.210184052587 Epoch: 97 -> Test Accuracy: 88.2 [98, 60] loss: 0.109 [98, 120] loss: 0.115 [98, 180] loss: 0.116 [98, 240] loss: 0.127 [98, 300] loss: 0.124 [98, 360] loss: 0.137 Epoch: 98 -> Loss: 0.177640527487 Epoch: 98 -> Test Accuracy: 88.43 [99, 60] loss: 0.118 [99, 120] loss: 0.105 [99, 180] loss: 0.123 [99, 240] loss: 0.127 [99, 300] loss: 0.134 [99, 360] loss: 0.144 Epoch: 99 -> Loss: 0.103200897574 Epoch: 99 -> Test Accuracy: 88.29 [100, 60] loss: 0.107 [100, 120] loss: 0.122 [100, 180] loss: 0.119 [100, 240] loss: 0.125 [100, 300] loss: 0.131 [100, 360] loss: 0.140 Epoch: 100 -> Loss: 0.155915349722 Epoch: 100 -> Test Accuracy: 88.03 [101, 60] loss: 0.123 [101, 120] loss: 0.131 [101, 180] loss: 0.131 [101, 240] loss: 0.125 [101, 300] loss: 0.131 [101, 360] loss: 0.165 Epoch: 101 -> Loss: 0.110289469361 Epoch: 101 -> Test Accuracy: 88.01 [102, 60] loss: 0.115 [102, 120] loss: 0.118 [102, 180] loss: 0.120 [102, 240] loss: 0.133 [102, 300] loss: 0.137 [102, 360] loss: 0.136 Epoch: 102 -> Loss: 0.199165135622 Epoch: 102 -> Test Accuracy: 88.09 [103, 60] loss: 0.103 [103, 120] loss: 0.120 [103, 180] loss: 0.130 [103, 240] loss: 0.136 [103, 300] loss: 0.129 [103, 360] loss: 0.131 Epoch: 103 -> Loss: 0.169885784388 Epoch: 103 -> Test Accuracy: 88.94 [104, 60] loss: 0.115 [104, 120] loss: 0.121 [104, 180] loss: 0.128 [104, 240] loss: 0.120 [104, 300] loss: 0.126 [104, 360] loss: 0.130 Epoch: 104 -> Loss: 0.119063951075 Epoch: 104 -> Test Accuracy: 88.54 [105, 60] loss: 0.115 [105, 120] loss: 0.121 [105, 180] loss: 0.123 [105, 240] loss: 0.124 [105, 300] loss: 0.127 [105, 360] loss: 0.135 Epoch: 105 -> Loss: 0.120506562293 Epoch: 105 -> Test Accuracy: 88.43 [106, 60] loss: 0.123 [106, 120] loss: 0.112 [106, 180] loss: 0.117 [106, 240] loss: 0.134 [106, 300] loss: 0.115 [106, 360] loss: 0.131 Epoch: 106 -> Loss: 0.187677174807 Epoch: 106 -> Test Accuracy: 88.6 [107, 60] loss: 0.112 [107, 120] loss: 0.114 [107, 180] loss: 0.127 [107, 240] loss: 0.118 [107, 300] loss: 0.138 [107, 360] loss: 0.133 Epoch: 107 -> Loss: 0.0318616889417 Epoch: 107 -> Test Accuracy: 87.91 [108, 60] loss: 0.105 [108, 120] loss: 0.111 [108, 180] loss: 0.112 [108, 240] loss: 0.131 [108, 300] loss: 0.140 [108, 360] loss: 0.138 Epoch: 108 -> Loss: 0.160463631153 Epoch: 108 -> Test Accuracy: 88.21 [109, 60] loss: 0.125 [109, 120] loss: 0.117 [109, 180] loss: 0.120 [109, 240] loss: 0.115 [109, 300] loss: 0.119 [109, 360] loss: 0.138 Epoch: 109 -> Loss: 0.112678147852 Epoch: 109 -> Test Accuracy: 88.54 [110, 60] loss: 0.114 [110, 120] loss: 0.103 [110, 180] loss: 0.112 [110, 240] loss: 0.125 [110, 300] loss: 0.137 [110, 360] loss: 0.136 Epoch: 110 -> Loss: 0.152882963419 Epoch: 110 -> Test Accuracy: 89.04 [111, 60] loss: 0.118 [111, 120] loss: 0.113 [111, 180] loss: 0.120 [111, 240] loss: 0.125 [111, 300] loss: 0.121 [111, 360] loss: 0.118 Epoch: 111 -> Loss: 0.0996031239629 Epoch: 111 -> Test Accuracy: 87.94 [112, 60] loss: 0.115 [112, 120] loss: 0.117 [112, 180] loss: 0.127 [112, 240] loss: 0.123 [112, 300] loss: 0.128 [112, 360] loss: 0.133 Epoch: 112 -> Loss: 0.166124328971 Epoch: 112 -> Test Accuracy: 87.92 [113, 60] loss: 0.107 [113, 120] loss: 0.117 [113, 180] loss: 0.114 [113, 240] loss: 0.117 [113, 300] loss: 0.119 [113, 360] loss: 0.135 Epoch: 113 -> Loss: 0.140588134527 Epoch: 113 -> Test Accuracy: 88.48 [114, 60] loss: 0.107 [114, 120] loss: 0.114 [114, 180] loss: 0.111 [114, 240] loss: 0.121 [114, 300] loss: 0.130 [114, 360] loss: 0.128 Epoch: 114 -> Loss: 0.212688833475 Epoch: 114 -> Test Accuracy: 88.04 [115, 60] loss: 0.108 [115, 120] loss: 0.117 [115, 180] loss: 0.123 [115, 240] loss: 0.130 [115, 300] loss: 0.123 [115, 360] loss: 0.138 Epoch: 115 -> Loss: 0.135155707598 Epoch: 115 -> Test Accuracy: 87.29 [116, 60] loss: 0.119 [116, 120] loss: 0.105 [116, 180] loss: 0.111 [116, 240] loss: 0.118 [116, 300] loss: 0.125 [116, 360] loss: 0.126 Epoch: 116 -> Loss: 0.0527365282178 Epoch: 116 -> Test Accuracy: 88.56 [117, 60] loss: 0.107 [117, 120] loss: 0.105 [117, 180] loss: 0.117 [117, 240] loss: 0.117 [117, 300] loss: 0.129 [117, 360] loss: 0.120 Epoch: 117 -> Loss: 0.0917150899768 Epoch: 117 -> Test Accuracy: 88.36 [118, 60] loss: 0.097 [118, 120] loss: 0.111 [118, 180] loss: 0.117 [118, 240] loss: 0.118 [118, 300] loss: 0.138 [118, 360] loss: 0.146 Epoch: 118 -> Loss: 0.15217718482 Epoch: 118 -> Test Accuracy: 87.54 [119, 60] loss: 0.115 [119, 120] loss: 0.114 [119, 180] loss: 0.115 [119, 240] loss: 0.118 [119, 300] loss: 0.118 [119, 360] loss: 0.132 Epoch: 119 -> Loss: 0.176103949547 Epoch: 119 -> Test Accuracy: 88.85 [120, 60] loss: 0.109 [120, 120] loss: 0.114 [120, 180] loss: 0.112 [120, 240] loss: 0.122 [120, 300] loss: 0.133 [120, 360] loss: 0.116 Epoch: 120 -> Loss: 0.0617025382817 Epoch: 120 -> Test Accuracy: 88.76 [121, 60] loss: 0.070 [121, 120] loss: 0.050 [121, 180] loss: 0.046 [121, 240] loss: 0.047 [121, 300] loss: 0.040 [121, 360] loss: 0.041 Epoch: 121 -> Loss: 0.0158089995384 Epoch: 121 -> Test Accuracy: 91.12 [122, 60] loss: 0.034 [122, 120] loss: 0.029 [122, 180] loss: 0.032 [122, 240] loss: 0.032 [122, 300] loss: 0.029 [122, 360] loss: 0.034 Epoch: 122 -> Loss: 0.0332125239074 Epoch: 122 -> Test Accuracy: 91.09 [123, 60] loss: 0.028 [123, 120] loss: 0.024 [123, 180] loss: 0.027 [123, 240] loss: 0.025 [123, 300] loss: 0.027 [123, 360] loss: 0.025 Epoch: 123 -> Loss: 0.0342334732413 Epoch: 123 -> Test Accuracy: 91.33 [124, 60] loss: 0.024 [124, 120] loss: 0.022 [124, 180] loss: 0.025 [124, 240] loss: 0.024 [124, 300] loss: 0.024 [124, 360] loss: 0.023 Epoch: 124 -> Loss: 0.0307826045901 Epoch: 124 -> Test Accuracy: 91.4 [125, 60] loss: 0.021 [125, 120] loss: 0.022 [125, 180] loss: 0.021 [125, 240] loss: 0.021 [125, 300] loss: 0.020 [125, 360] loss: 0.021 Epoch: 125 -> Loss: 0.00836864393204 Epoch: 125 -> Test Accuracy: 91.36 [126, 60] loss: 0.020 [126, 120] loss: 0.018 [126, 180] loss: 0.018 [126, 240] loss: 0.020 [126, 300] loss: 0.020 [126, 360] loss: 0.020 Epoch: 126 -> Loss: 0.018958395347 Epoch: 126 -> Test Accuracy: 91.32 [127, 60] loss: 0.016 [127, 120] loss: 0.017 [127, 180] loss: 0.019 [127, 240] loss: 0.019 [127, 300] loss: 0.020 [127, 360] loss: 0.019 Epoch: 127 -> Loss: 0.0133167505264 Epoch: 127 -> Test Accuracy: 91.38 [128, 60] loss: 0.017 [128, 120] loss: 0.014 [128, 180] loss: 0.016 [128, 240] loss: 0.016 [128, 300] loss: 0.017 [128, 360] loss: 0.017 Epoch: 128 -> Loss: 0.0116452155635 Epoch: 128 -> Test Accuracy: 91.46 [129, 60] loss: 0.015 [129, 120] loss: 0.015 [129, 180] loss: 0.015 [129, 240] loss: 0.014 [129, 300] loss: 0.016 [129, 360] loss: 0.018 Epoch: 129 -> Loss: 0.0130935786292 Epoch: 129 -> Test Accuracy: 91.05 [130, 60] loss: 0.015 [130, 120] loss: 0.013 [130, 180] loss: 0.015 [130, 240] loss: 0.015 [130, 300] loss: 0.015 [130, 360] loss: 0.016 Epoch: 130 -> Loss: 0.00634835381061 Epoch: 130 -> Test Accuracy: 91.38 [131, 60] loss: 0.014 [131, 120] loss: 0.015 [131, 180] loss: 0.014 [131, 240] loss: 0.016 [131, 300] loss: 0.016 [131, 360] loss: 0.014 Epoch: 131 -> Loss: 0.021570796147 Epoch: 131 -> Test Accuracy: 91.49 [132, 60] loss: 0.013 [132, 120] loss: 0.014 [132, 180] loss: 0.014 [132, 240] loss: 0.015 [132, 300] loss: 0.014 [132, 360] loss: 0.013 Epoch: 132 -> Loss: 0.00960294343531 Epoch: 132 -> Test Accuracy: 91.59 [133, 60] loss: 0.013 [133, 120] loss: 0.015 [133, 180] loss: 0.015 [133, 240] loss: 0.014 [133, 300] loss: 0.015 [133, 360] loss: 0.015 Epoch: 133 -> Loss: 0.0142041146755 Epoch: 133 -> Test Accuracy: 91.4 [134, 60] loss: 0.012 [134, 120] loss: 0.012 [134, 180] loss: 0.014 [134, 240] loss: 0.013 [134, 300] loss: 0.013 [134, 360] loss: 0.013 Epoch: 134 -> Loss: 0.00717277545482 Epoch: 134 -> Test Accuracy: 91.26 [135, 60] loss: 0.014 [135, 120] loss: 0.014 [135, 180] loss: 0.014 [135, 240] loss: 0.012 [135, 300] loss: 0.012 [135, 360] loss: 0.012 Epoch: 135 -> Loss: 0.0177748799324 Epoch: 135 -> Test Accuracy: 91.39 [136, 60] loss: 0.011 [136, 120] loss: 0.013 [136, 180] loss: 0.013 [136, 240] loss: 0.015 [136, 300] loss: 0.012 [136, 360] loss: 0.012 Epoch: 136 -> Loss: 0.00705651659518 Epoch: 136 -> Test Accuracy: 91.33 [137, 60] loss: 0.011 [137, 120] loss: 0.014 [137, 180] loss: 0.013 [137, 240] loss: 0.013 [137, 300] loss: 0.012 [137, 360] loss: 0.012 Epoch: 137 -> Loss: 0.0143475178629 Epoch: 137 -> Test Accuracy: 91.21 [138, 60] loss: 0.011 [138, 120] loss: 0.012 [138, 180] loss: 0.013 [138, 240] loss: 0.012 [138, 300] loss: 0.012 [138, 360] loss: 0.013 Epoch: 138 -> Loss: 0.0162122733891 Epoch: 138 -> Test Accuracy: 91.4 [139, 60] loss: 0.011 [139, 120] loss: 0.011 [139, 180] loss: 0.011 [139, 240] loss: 0.012 [139, 300] loss: 0.012 [139, 360] loss: 0.011 Epoch: 139 -> Loss: 0.01758239232 Epoch: 139 -> Test Accuracy: 91.37 [140, 60] loss: 0.010 [140, 120] loss: 0.011 [140, 180] loss: 0.011 [140, 240] loss: 0.012 [140, 300] loss: 0.010 [140, 360] loss: 0.012 Epoch: 140 -> Loss: 0.0165322963148 Epoch: 140 -> Test Accuracy: 91.22 [141, 60] loss: 0.011 [141, 120] loss: 0.011 [141, 180] loss: 0.011 [141, 240] loss: 0.012 [141, 300] loss: 0.010 [141, 360] loss: 0.010 Epoch: 141 -> Loss: 0.00455831876025 Epoch: 141 -> Test Accuracy: 91.24 [142, 60] loss: 0.010 [142, 120] loss: 0.011 [142, 180] loss: 0.010 [142, 240] loss: 0.011 [142, 300] loss: 0.011 [142, 360] loss: 0.012 Epoch: 142 -> Loss: 0.0214500371367 Epoch: 142 -> Test Accuracy: 91.4 [143, 60] loss: 0.010 [143, 120] loss: 0.009 [143, 180] loss: 0.010 [143, 240] loss: 0.010 [143, 300] loss: 0.011 [143, 360] loss: 0.011 Epoch: 143 -> Loss: 0.0189341306686 Epoch: 143 -> Test Accuracy: 91.35 [144, 60] loss: 0.010 [144, 120] loss: 0.010 [144, 180] loss: 0.011 [144, 240] loss: 0.011 [144, 300] loss: 0.010 [144, 360] loss: 0.011 Epoch: 144 -> Loss: 0.0113361775875 Epoch: 144 -> Test Accuracy: 91.29 [145, 60] loss: 0.011 [145, 120] loss: 0.010 [145, 180] loss: 0.009 [145, 240] loss: 0.011 [145, 300] loss: 0.010 [145, 360] loss: 0.010 Epoch: 145 -> Loss: 0.00607524532825 Epoch: 145 -> Test Accuracy: 91.31 [146, 60] loss: 0.010 [146, 120] loss: 0.011 [146, 180] loss: 0.010 [146, 240] loss: 0.010 [146, 300] loss: 0.010 [146, 360] loss: 0.010 Epoch: 146 -> Loss: 0.00923008285463 Epoch: 146 -> Test Accuracy: 91.2 [147, 60] loss: 0.009 [147, 120] loss: 0.010 [147, 180] loss: 0.011 [147, 240] loss: 0.011 [147, 300] loss: 0.010 [147, 360] loss: 0.011 Epoch: 147 -> Loss: 0.00989549141377 Epoch: 147 -> Test Accuracy: 91.22 [148, 60] loss: 0.009 [148, 120] loss: 0.010 [148, 180] loss: 0.010 [148, 240] loss: 0.011 [148, 300] loss: 0.011 [148, 360] loss: 0.011 Epoch: 148 -> Loss: 0.00719866156578 Epoch: 148 -> Test Accuracy: 91.43 [149, 60] loss: 0.011 [149, 120] loss: 0.010 [149, 180] loss: 0.010 [149, 240] loss: 0.010 [149, 300] loss: 0.009 [149, 360] loss: 0.010 Epoch: 149 -> Loss: 0.0087026655674 Epoch: 149 -> Test Accuracy: 91.37 [150, 60] loss: 0.009 [150, 120] loss: 0.010 [150, 180] loss: 0.010 [150, 240] loss: 0.011 [150, 300] loss: 0.011 [150, 360] loss: 0.011 Epoch: 150 -> Loss: 0.0234594587237 Epoch: 150 -> Test Accuracy: 91.19 [151, 60] loss: 0.010 [151, 120] loss: 0.011 [151, 180] loss: 0.010 [151, 240] loss: 0.010 [151, 300] loss: 0.010 [151, 360] loss: 0.011 Epoch: 151 -> Loss: 0.0275905188173 Epoch: 151 -> Test Accuracy: 91.08 [152, 60] loss: 0.011 [152, 120] loss: 0.010 [152, 180] loss: 0.011 [152, 240] loss: 0.011 [152, 300] loss: 0.010 [152, 360] loss: 0.010 Epoch: 152 -> Loss: 0.0117717506364 Epoch: 152 -> Test Accuracy: 91.34 [153, 60] loss: 0.009 [153, 120] loss: 0.008 [153, 180] loss: 0.010 [153, 240] loss: 0.009 [153, 300] loss: 0.010 [153, 360] loss: 0.010 Epoch: 153 -> Loss: 0.0202627424151 Epoch: 153 -> Test Accuracy: 91.17 [154, 60] loss: 0.009 [154, 120] loss: 0.010 [154, 180] loss: 0.009 [154, 240] loss: 0.010 [154, 300] loss: 0.010 [154, 360] loss: 0.012 Epoch: 154 -> Loss: 0.00777658214793 Epoch: 154 -> Test Accuracy: 91.19 [155, 60] loss: 0.009 [155, 120] loss: 0.010 [155, 180] loss: 0.010 [155, 240] loss: 0.010 [155, 300] loss: 0.009 [155, 360] loss: 0.009 Epoch: 155 -> Loss: 0.00603697309271 Epoch: 155 -> Test Accuracy: 91.48 [156, 60] loss: 0.009 [156, 120] loss: 0.010 [156, 180] loss: 0.009 [156, 240] loss: 0.009 [156, 300] loss: 0.010 [156, 360] loss: 0.009 Epoch: 156 -> Loss: 0.0179765280336 Epoch: 156 -> Test Accuracy: 91.16 [157, 60] loss: 0.010 [157, 120] loss: 0.010 [157, 180] loss: 0.010 [157, 240] loss: 0.010 [157, 300] loss: 0.009 [157, 360] loss: 0.009 Epoch: 157 -> Loss: 0.0177674647421 Epoch: 157 -> Test Accuracy: 91.23 [158, 60] loss: 0.010 [158, 120] loss: 0.009 [158, 180] loss: 0.010 [158, 240] loss: 0.010 [158, 300] loss: 0.009 [158, 360] loss: 0.010 Epoch: 158 -> Loss: 0.00916373729706 Epoch: 158 -> Test Accuracy: 91.33 [159, 60] loss: 0.009 [159, 120] loss: 0.009 [159, 180] loss: 0.010 [159, 240] loss: 0.008 [159, 300] loss: 0.009 [159, 360] loss: 0.009 Epoch: 159 -> Loss: 0.0119180288166 Epoch: 159 -> Test Accuracy: 91.17 [160, 60] loss: 0.008 [160, 120] loss: 0.008 [160, 180] loss: 0.008 [160, 240] loss: 0.010 [160, 300] loss: 0.009 [160, 360] loss: 0.009 Epoch: 160 -> Loss: 0.0299162622541 Epoch: 160 -> Test Accuracy: 91.1 [161, 60] loss: 0.008 [161, 120] loss: 0.009 [161, 180] loss: 0.009 [161, 240] loss: 0.008 [161, 300] loss: 0.008 [161, 360] loss: 0.009 Epoch: 161 -> Loss: 0.00964940153062 Epoch: 161 -> Test Accuracy: 91.13 [162, 60] loss: 0.008 [162, 120] loss: 0.007 [162, 180] loss: 0.009 [162, 240] loss: 0.008 [162, 300] loss: 0.007 [162, 360] loss: 0.008 Epoch: 162 -> Loss: 0.00648107519373 Epoch: 162 -> Test Accuracy: 91.25 [163, 60] loss: 0.007 [163, 120] loss: 0.008 [163, 180] loss: 0.008 [163, 240] loss: 0.008 [163, 300] loss: 0.008 [163, 360] loss: 0.008 Epoch: 163 -> Loss: 0.0105269672349 Epoch: 163 -> Test Accuracy: 91.33 [164, 60] loss: 0.008 [164, 120] loss: 0.007 [164, 180] loss: 0.007 [164, 240] loss: 0.007 [164, 300] loss: 0.007 [164, 360] loss: 0.008 Epoch: 164 -> Loss: 0.0457272157073 Epoch: 164 -> Test Accuracy: 91.28 [165, 60] loss: 0.008 [165, 120] loss: 0.007 [165, 180] loss: 0.007 [165, 240] loss: 0.008 [165, 300] loss: 0.007 [165, 360] loss: 0.007 Epoch: 165 -> Loss: 0.0171728171408 Epoch: 165 -> Test Accuracy: 91.4 [166, 60] loss: 0.007 [166, 120] loss: 0.008 [166, 180] loss: 0.007 [166, 240] loss: 0.007 [166, 300] loss: 0.008 [166, 360] loss: 0.008 Epoch: 166 -> Loss: 0.0133938370273 Epoch: 166 -> Test Accuracy: 91.4 [167, 60] loss: 0.007 [167, 120] loss: 0.007 [167, 180] loss: 0.007 [167, 240] loss: 0.007 [167, 300] loss: 0.007 [167, 360] loss: 0.006 Epoch: 167 -> Loss: 0.0162021405995 Epoch: 167 -> Test Accuracy: 91.4 [168, 60] loss: 0.007 [168, 120] loss: 0.007 [168, 180] loss: 0.006 [168, 240] loss: 0.008 [168, 300] loss: 0.007 [168, 360] loss: 0.007 Epoch: 168 -> Loss: 0.00573644647375 Epoch: 168 -> Test Accuracy: 91.37 [169, 60] loss: 0.007 [169, 120] loss: 0.007 [169, 180] loss: 0.006 [169, 240] loss: 0.007 [169, 300] loss: 0.007 [169, 360] loss: 0.007 Epoch: 169 -> Loss: 0.00966047029942 Epoch: 169 -> Test Accuracy: 91.4 [170, 60] loss: 0.007 [170, 120] loss: 0.007 [170, 180] loss: 0.007 [170, 240] loss: 0.008 [170, 300] loss: 0.007 [170, 360] loss: 0.007 Epoch: 170 -> Loss: 0.0125759299845 Epoch: 170 -> Test Accuracy: 91.34 [171, 60] loss: 0.006 [171, 120] loss: 0.006 [171, 180] loss: 0.007 [171, 240] loss: 0.007 [171, 300] loss: 0.007 [171, 360] loss: 0.007 Epoch: 171 -> Loss: 0.00646898150444 Epoch: 171 -> Test Accuracy: 91.26 [172, 60] loss: 0.007 [172, 120] loss: 0.007 [172, 180] loss: 0.007 [172, 240] loss: 0.007 [172, 300] loss: 0.006 [172, 360] loss: 0.007 Epoch: 172 -> Loss: 0.0301642324775 Epoch: 172 -> Test Accuracy: 91.3 [173, 60] loss: 0.006 [173, 120] loss: 0.007 [173, 180] loss: 0.007 [173, 240] loss: 0.007 [173, 300] loss: 0.007 [173, 360] loss: 0.007 Epoch: 173 -> Loss: 0.00349472160451 Epoch: 173 -> Test Accuracy: 91.33 [174, 60] loss: 0.007 [174, 120] loss: 0.007 [174, 180] loss: 0.007 [174, 240] loss: 0.007 [174, 300] loss: 0.007 [174, 360] loss: 0.008 Epoch: 174 -> Loss: 0.00375400180928 Epoch: 174 -> Test Accuracy: 91.36 [175, 60] loss: 0.007 [175, 120] loss: 0.007 [175, 180] loss: 0.007 [175, 240] loss: 0.007 [175, 300] loss: 0.007 [175, 360] loss: 0.007 Epoch: 175 -> Loss: 0.0120144542307 Epoch: 175 -> Test Accuracy: 91.42 [176, 60] loss: 0.007 [176, 120] loss: 0.006 [176, 180] loss: 0.008 [176, 240] loss: 0.006 [176, 300] loss: 0.007 [176, 360] loss: 0.007 Epoch: 176 -> Loss: 0.0169802550226 Epoch: 176 -> Test Accuracy: 91.35 [177, 60] loss: 0.007 [177, 120] loss: 0.007 [177, 180] loss: 0.007 [177, 240] loss: 0.007 [177, 300] loss: 0.007 [177, 360] loss: 0.006 Epoch: 177 -> Loss: 0.00630267243832 Epoch: 177 -> Test Accuracy: 91.38 [178, 60] loss: 0.007 [178, 120] loss: 0.007 [178, 180] loss: 0.007 [178, 240] loss: 0.007 [178, 300] loss: 0.006 [178, 360] loss: 0.007 Epoch: 178 -> Loss: 0.0068407417275 Epoch: 178 -> Test Accuracy: 91.21 [179, 60] loss: 0.007 [179, 120] loss: 0.007 [179, 180] loss: 0.007 [179, 240] loss: 0.007 [179, 300] loss: 0.007 [179, 360] loss: 0.007 Epoch: 179 -> Loss: 0.01610468328 Epoch: 179 -> Test Accuracy: 91.35 [180, 60] loss: 0.007 [180, 120] loss: 0.006 [180, 180] loss: 0.007 [180, 240] loss: 0.006 [180, 300] loss: 0.007 [180, 360] loss: 0.008 Epoch: 180 -> Loss: 0.00931457895786 Epoch: 180 -> Test Accuracy: 91.29 [181, 60] loss: 0.007 [181, 120] loss: 0.007 [181, 180] loss: 0.007 [181, 240] loss: 0.007 [181, 300] loss: 0.007 [181, 360] loss: 0.007 Epoch: 181 -> Loss: 0.00952905416489 Epoch: 181 -> Test Accuracy: 91.34 [182, 60] loss: 0.006 [182, 120] loss: 0.007 [182, 180] loss: 0.007 [182, 240] loss: 0.006 [182, 300] loss: 0.008 [182, 360] loss: 0.006 Epoch: 182 -> Loss: 0.00699643511325 Epoch: 182 -> Test Accuracy: 91.43 [183, 60] loss: 0.007 [183, 120] loss: 0.007 [183, 180] loss: 0.007 [183, 240] loss: 0.008 [183, 300] loss: 0.006 [183, 360] loss: 0.006 Epoch: 183 -> Loss: 0.00621772417799 Epoch: 183 -> Test Accuracy: 91.32 [184, 60] loss: 0.007 [184, 120] loss: 0.007 [184, 180] loss: 0.006 [184, 240] loss: 0.007 [184, 300] loss: 0.007 [184, 360] loss: 0.007 Epoch: 184 -> Loss: 0.00833785533905 Epoch: 184 -> Test Accuracy: 91.33 [185, 60] loss: 0.007 [185, 120] loss: 0.007 [185, 180] loss: 0.008 [185, 240] loss: 0.007 [185, 300] loss: 0.006 [185, 360] loss: 0.006 Epoch: 185 -> Loss: 0.00589191913605 Epoch: 185 -> Test Accuracy: 91.3 [186, 60] loss: 0.007 [186, 120] loss: 0.008 [186, 180] loss: 0.007 [186, 240] loss: 0.007 [186, 300] loss: 0.007 [186, 360] loss: 0.008 Epoch: 186 -> Loss: 0.00488076219335 Epoch: 186 -> Test Accuracy: 91.36 [187, 60] loss: 0.006 [187, 120] loss: 0.007 [187, 180] loss: 0.006 [187, 240] loss: 0.007 [187, 300] loss: 0.007 [187, 360] loss: 0.007 Epoch: 187 -> Loss: 0.0116668641567 Epoch: 187 -> Test Accuracy: 91.44 [188, 60] loss: 0.007 [188, 120] loss: 0.007 [188, 180] loss: 0.007 [188, 240] loss: 0.007 [188, 300] loss: 0.007 [188, 360] loss: 0.006 Epoch: 188 -> Loss: 0.00331798801199 Epoch: 188 -> Test Accuracy: 91.47 [189, 60] loss: 0.006 [189, 120] loss: 0.006 [189, 180] loss: 0.007 [189, 240] loss: 0.007 [189, 300] loss: 0.007 [189, 360] loss: 0.007 Epoch: 189 -> Loss: 0.00637602806091 Epoch: 189 -> Test Accuracy: 91.4 [190, 60] loss: 0.007 [190, 120] loss: 0.007 [190, 180] loss: 0.007 [190, 240] loss: 0.007 [190, 300] loss: 0.006 [190, 360] loss: 0.007 Epoch: 190 -> Loss: 0.0175461675972 Epoch: 190 -> Test Accuracy: 91.38 [191, 60] loss: 0.007 [191, 120] loss: 0.007 [191, 180] loss: 0.007 [191, 240] loss: 0.007 [191, 300] loss: 0.006 [191, 360] loss: 0.006 Epoch: 191 -> Loss: 0.00864890217781 Epoch: 191 -> Test Accuracy: 91.42 [192, 60] loss: 0.007 [192, 120] loss: 0.007 [192, 180] loss: 0.006 [192, 240] loss: 0.006 [192, 300] loss: 0.007 [192, 360] loss: 0.006 Epoch: 192 -> Loss: 0.003874379443 Epoch: 192 -> Test Accuracy: 91.48 [193, 60] loss: 0.007 [193, 120] loss: 0.006 [193, 180] loss: 0.006 [193, 240] loss: 0.007 [193, 300] loss: 0.007 [193, 360] loss: 0.008 Epoch: 193 -> Loss: 0.00823497213423 Epoch: 193 -> Test Accuracy: 91.44 [194, 60] loss: 0.007 [194, 120] loss: 0.006 [194, 180] loss: 0.006 [194, 240] loss: 0.007 [194, 300] loss: 0.007 [194, 360] loss: 0.007 Epoch: 194 -> Loss: 0.0043452619575 Epoch: 194 -> Test Accuracy: 91.34 [195, 60] loss: 0.006 [195, 120] loss: 0.006 [195, 180] loss: 0.007 [195, 240] loss: 0.007 [195, 300] loss: 0.006 [195, 360] loss: 0.007 Epoch: 195 -> Loss: 0.0079355482012 Epoch: 195 -> Test Accuracy: 91.53 [196, 60] loss: 0.006 [196, 120] loss: 0.006 [196, 180] loss: 0.006 [196, 240] loss: 0.006 [196, 300] loss: 0.007 [196, 360] loss: 0.006 Epoch: 196 -> Loss: 0.0104473233223 Epoch: 196 -> Test Accuracy: 91.46 [197, 60] loss: 0.007 [197, 120] loss: 0.006 [197, 180] loss: 0.007 [197, 240] loss: 0.007 [197, 300] loss: 0.007 [197, 360] loss: 0.007 Epoch: 197 -> Loss: 0.00406329613179 Epoch: 197 -> Test Accuracy: 91.51 [198, 60] loss: 0.007 [198, 120] loss: 0.007 [198, 180] loss: 0.006 [198, 240] loss: 0.007 [198, 300] loss: 0.007 [198, 360] loss: 0.007 Epoch: 198 -> Loss: 0.0194802395999 Epoch: 198 -> Test Accuracy: 91.48 [199, 60] loss: 0.007 [199, 120] loss: 0.006 [199, 180] loss: 0.007 [199, 240] loss: 0.006 [199, 300] loss: 0.007 [199, 360] loss: 0.007 Epoch: 199 -> Loss: 0.00719060888514 Epoch: 199 -> Test Accuracy: 91.49 [200, 60] loss: 0.007 [200, 120] loss: 0.006 [200, 180] loss: 0.006 [200, 240] loss: 0.007 [200, 300] loss: 0.007 [200, 360] loss: 0.006 Epoch: 200 -> Loss: 0.00321622495539 Epoch: 200 -> Test Accuracy: 91.37 Finished Training
In [14]:
# save variables
fm.save_variable([semi_loss_log, semi_accuracy_log, super_loss_log, super_accuracy_log], "semi-supervised")
Evaluate Test Accuracies¶
In [6]:
# 3 ConvBlock RotNet model and Classifiers
ev.evaluate_all(3, testloader, classes)
Evaluating RotNet model with 3 Convolutional Blocks: Evaluating Rotation Task: Test Accuracy: 92.190 % Accuracy per class: Test Accuracy of original : 92.100 % Test Accuracy of 90 rotation : 92.110 % Test Accuracy of 180 rotation : 92.140 % Test Accuracy of 270 rotation : 92.410 % -------------------------------------------------------------------------------- Starting to evaluate Non-Linear Classifier: Evaluating Non-Linear Classifier on Convolutional Block 1: Test Accuracy: 83.280 % Accuracy per class: Test Accuracy of plane : 85.100 % Test Accuracy of car : 91.400 % Test Accuracy of bird : 76.900 % Test Accuracy of cat : 69.000 % Test Accuracy of deer : 79.700 % Test Accuracy of dog : 72.900 % Test Accuracy of frog : 89.500 % Test Accuracy of horse : 86.300 % Test Accuracy of ship : 91.100 % Test Accuracy of truck : 90.900 % Evaluating Non-Linear Classifier on Convolutional Block 2: Test Accuracy: 86.510 % Accuracy per class: Test Accuracy of plane : 87.300 % Test Accuracy of car : 92.600 % Test Accuracy of bird : 82.400 % Test Accuracy of cat : 77.100 % Test Accuracy of deer : 86.600 % Test Accuracy of dog : 77.600 % Test Accuracy of frog : 90.400 % Test Accuracy of horse : 88.600 % Test Accuracy of ship : 92.400 % Test Accuracy of truck : 90.100 % Evaluating Non-Linear Classifier on Convolutional Block 3: Test Accuracy: 54.070 % Accuracy per class: Test Accuracy of plane : 61.800 % Test Accuracy of car : 59.200 % Test Accuracy of bird : 43.100 % Test Accuracy of cat : 36.500 % Test Accuracy of deer : 50.300 % Test Accuracy of dog : 51.100 % Test Accuracy of frog : 61.300 % Test Accuracy of horse : 54.600 % Test Accuracy of ship : 60.000 % Test Accuracy of truck : 62.800 % -------------------------------------------------------------------------------- Starting to evaluate Convolutional Classifier: Evaluating Convolutional Classifier on Convolutional Block 1 Test Accuracy: 86.850 % Accuracy per class: Test Accuracy of plane : 88.300 % Test Accuracy of car : 93.900 % Test Accuracy of bird : 80.200 % Test Accuracy of cat : 76.900 % Test Accuracy of deer : 86.700 % Test Accuracy of dog : 78.400 % Test Accuracy of frog : 92.500 % Test Accuracy of horse : 89.200 % Test Accuracy of ship : 91.500 % Test Accuracy of truck : 90.900 % Evaluating Convolutional Classifier on Convolutional Block 2 Test Accuracy: 88.820 % Accuracy per class: Test Accuracy of plane : 89.200 % Test Accuracy of car : 94.000 % Test Accuracy of bird : 84.400 % Test Accuracy of cat : 79.900 % Test Accuracy of deer : 89.800 % Test Accuracy of dog : 83.400 % Test Accuracy of frog : 92.200 % Test Accuracy of horse : 90.000 % Test Accuracy of ship : 93.300 % Test Accuracy of truck : 92.000 % Evaluating Convolutional Classifier on Convolutional Block 3 Test Accuracy: 61.910 % Accuracy per class: Test Accuracy of plane : 64.700 % Test Accuracy of car : 64.500 % Test Accuracy of bird : 51.100 % Test Accuracy of cat : 49.300 % Test Accuracy of deer : 61.200 % Test Accuracy of dog : 55.300 % Test Accuracy of frog : 71.600 % Test Accuracy of horse : 61.400 % Test Accuracy of ship : 67.900 % Test Accuracy of truck : 72.100 %
Out[6]:
{'Accuracy ConvClassifier ConvBlock 1': 86.85,
'Accuracy ConvClassifier ConvBlock 2': 88.82,
'Accuracy ConvClassifier ConvBlock 3': 61.91,
'Accuracy Non-Linear ConvBlock 1': 83.28,
'Accuracy Non-Linear ConvBlock 2': 86.51,
'Accuracy Non-Linear ConvBlock 3': 54.07,
'Accuracy Rotation Task': 92.19,
'Class Accuracy ConvClassifier ConvBlock 1': [88.3,
93.9,
80.2,
76.9,
86.7,
78.4,
92.5,
89.2,
91.5,
90.9],
'Class Accuracy ConvClassifier ConvBlock 2': [89.2,
94.0,
84.4,
79.9,
89.8,
83.4,
92.2,
90.0,
93.3,
92.0],
'Class Accuracy ConvClassifier ConvBlock 3': [64.7,
64.5,
51.1,
49.3,
61.2,
55.3,
71.6,
61.4,
67.9,
72.1],
'Class Accuracy Non-Linear ConvBlock 1': [85.1,
91.4,
76.9,
69.0,
79.7,
72.9,
89.5,
86.3,
91.1,
90.9],
'Class Accuracy Non-Linear ConvBlock 2': [87.3,
92.6,
82.4,
77.1,
86.6,
77.6,
90.4,
88.6,
92.4,
90.1],
'Class Accuracy Non-Linear ConvBlock 3': [61.8,
59.2,
43.1,
36.5,
50.3,
51.1,
61.3,
54.6,
60.0,
62.8],
'Class Accuracy Rotation Task': [92.1, 92.11, 92.14, 92.41]}
In [15]:
# 4 ConvBlock RotNet model and Classifiers
ev.evaluate_all(4, testloader, classes)
Evaluating RotNet model with 4 Convolutional Blocks: Evaluating Rotation Task: Test Accuracy: 92.632 % Accuracy per class: Test Accuracy of original : 92.560 % Test Accuracy of 90 rotation : 92.350 % Test Accuracy of 180 rotation : 92.880 % Test Accuracy of 270 rotation : 92.740 % -------------------------------------------------------------------------------- Starting to evaluate Non-Linear Classifier: Evaluating Non-Linear Classifier on Convolutional Block 1: Test Accuracy: 83.120 % Accuracy per class: Test Accuracy of plane : 84.200 % Test Accuracy of car : 91.800 % Test Accuracy of bird : 75.800 % Test Accuracy of cat : 67.300 % Test Accuracy of deer : 82.400 % Test Accuracy of dog : 74.100 % Test Accuracy of frog : 88.700 % Test Accuracy of horse : 86.600 % Test Accuracy of ship : 90.800 % Test Accuracy of truck : 89.500 % Evaluating Non-Linear Classifier on Convolutional Block 2: Test Accuracy: 86.600 % Accuracy per class: Test Accuracy of plane : 86.700 % Test Accuracy of car : 92.500 % Test Accuracy of bird : 82.700 % Test Accuracy of cat : 75.300 % Test Accuracy of deer : 85.300 % Test Accuracy of dog : 80.000 % Test Accuracy of frog : 91.300 % Test Accuracy of horse : 89.500 % Test Accuracy of ship : 91.200 % Test Accuracy of truck : 91.500 % Evaluating Non-Linear Classifier on Convolutional Block 3: Test Accuracy: 82.540 % Accuracy per class: Test Accuracy of plane : 82.400 % Test Accuracy of car : 90.000 % Test Accuracy of bird : 78.000 % Test Accuracy of cat : 72.000 % Test Accuracy of deer : 81.400 % Test Accuracy of dog : 74.300 % Test Accuracy of frog : 87.500 % Test Accuracy of horse : 86.600 % Test Accuracy of ship : 86.500 % Test Accuracy of truck : 86.700 % Evaluating Non-Linear Classifier on Convolutional Block 4: Test Accuracy: 45.300 % Accuracy per class: Test Accuracy of plane : 55.900 % Test Accuracy of car : 49.300 % Test Accuracy of bird : 34.200 % Test Accuracy of cat : 34.800 % Test Accuracy of deer : 34.200 % Test Accuracy of dog : 38.600 % Test Accuracy of frog : 56.200 % Test Accuracy of horse : 47.900 % Test Accuracy of ship : 51.800 % Test Accuracy of truck : 50.100 % -------------------------------------------------------------------------------- Starting to evaluate Convolutional Classifier: Evaluating Convolutional Classifier on Convolutional Block 1 Test Accuracy: 86.580 % Accuracy per class: Test Accuracy of plane : 89.300 % Test Accuracy of car : 93.700 % Test Accuracy of bird : 80.000 % Test Accuracy of cat : 75.200 % Test Accuracy of deer : 84.200 % Test Accuracy of dog : 79.200 % Test Accuracy of frog : 91.700 % Test Accuracy of horse : 88.700 % Test Accuracy of ship : 91.800 % Test Accuracy of truck : 92.000 % Evaluating Convolutional Classifier on Convolutional Block 2 Test Accuracy: 88.830 % Accuracy per class: Test Accuracy of plane : 88.400 % Test Accuracy of car : 93.100 % Test Accuracy of bird : 86.100 % Test Accuracy of cat : 78.900 % Test Accuracy of deer : 89.300 % Test Accuracy of dog : 84.500 % Test Accuracy of frog : 92.000 % Test Accuracy of horse : 91.300 % Test Accuracy of ship : 93.200 % Test Accuracy of truck : 91.500 % Evaluating Convolutional Classifier on Convolutional Block 3 Test Accuracy: 84.250 % Accuracy per class: Test Accuracy of plane : 85.000 % Test Accuracy of car : 90.400 % Test Accuracy of bird : 79.700 % Test Accuracy of cat : 68.800 % Test Accuracy of deer : 84.200 % Test Accuracy of dog : 80.100 % Test Accuracy of frog : 89.100 % Test Accuracy of horse : 87.800 % Test Accuracy of ship : 89.300 % Test Accuracy of truck : 88.100 % Evaluating Convolutional Classifier on Convolutional Block 4 Test Accuracy: 53.400 % Accuracy per class: Test Accuracy of plane : 62.600 % Test Accuracy of car : 57.600 % Test Accuracy of bird : 41.900 % Test Accuracy of cat : 35.100 % Test Accuracy of deer : 48.600 % Test Accuracy of dog : 49.300 % Test Accuracy of frog : 64.600 % Test Accuracy of horse : 55.400 % Test Accuracy of ship : 60.200 % Test Accuracy of truck : 58.700 %
Out[15]:
{'Accuracy ConvClassifier ConvBlock 1': 86.58,
'Accuracy ConvClassifier ConvBlock 2': 88.83,
'Accuracy ConvClassifier ConvBlock 3': 84.25,
'Accuracy ConvClassifier ConvBlock 4': 53.4,
'Accuracy Non-Linear ConvBlock 1': 83.12,
'Accuracy Non-Linear ConvBlock 2': 86.6,
'Accuracy Non-Linear ConvBlock 3': 82.54,
'Accuracy Non-Linear ConvBlock 4': 45.3,
'Accuracy Rotation Task': 92.6325,
'Class Accuracy ConvClassifier ConvBlock 1': [89.3,
93.7,
80.0,
75.2,
84.2,
79.2,
91.7,
88.7,
91.8,
92.0],
'Class Accuracy ConvClassifier ConvBlock 2': [88.4,
93.1,
86.1,
78.9,
89.3,
84.5,
92.0,
91.3,
93.2,
91.5],
'Class Accuracy ConvClassifier ConvBlock 3': [85.0,
90.4,
79.7,
68.8,
84.2,
80.1,
89.1,
87.8,
89.3,
88.1],
'Class Accuracy ConvClassifier ConvBlock 4': [62.6,
57.6,
41.9,
35.1,
48.6,
49.3,
64.6,
55.4,
60.2,
58.7],
'Class Accuracy Non-Linear ConvBlock 1': [84.2,
91.8,
75.8,
67.3,
82.4,
74.1,
88.7,
86.6,
90.8,
89.5],
'Class Accuracy Non-Linear ConvBlock 2': [86.7,
92.5,
82.7,
75.3,
85.3,
80.0,
91.3,
89.5,
91.2,
91.5],
'Class Accuracy Non-Linear ConvBlock 3': [82.4,
90.0,
78.0,
72.0,
81.4,
74.3,
87.5,
86.6,
86.5,
86.7],
'Class Accuracy Non-Linear ConvBlock 4': [55.9,
49.3,
34.2,
34.8,
34.2,
38.6,
56.2,
47.9,
51.8,
50.1],
'Class Accuracy Rotation Task': [92.56, 92.35, 92.88, 92.74]}
In [8]:
# 5 ConvBlock RotNet model and Classifiers
ev.evaluate_all(5, testloader, classes)
Evaluating RotNet model with 5 Convolutional Blocks: Evaluating Rotation Task: Test Accuracy: 92.088 % Accuracy per class: Test Accuracy of original : 92.380 % Test Accuracy of 90 rotation : 92.150 % Test Accuracy of 180 rotation : 91.850 % Test Accuracy of 270 rotation : 91.970 % -------------------------------------------------------------------------------- Starting to evaluate Non-Linear Classifier: Evaluating Non-Linear Classifier on Convolutional Block 1: Test Accuracy: 82.990 % Accuracy per class: Test Accuracy of plane : 82.700 % Test Accuracy of car : 90.200 % Test Accuracy of bird : 77.800 % Test Accuracy of cat : 68.300 % Test Accuracy of deer : 79.100 % Test Accuracy of dog : 73.500 % Test Accuracy of frog : 88.600 % Test Accuracy of horse : 88.000 % Test Accuracy of ship : 91.800 % Test Accuracy of truck : 89.900 % Evaluating Non-Linear Classifier on Convolutional Block 2: Test Accuracy: 86.610 % Accuracy per class: Test Accuracy of plane : 88.900 % Test Accuracy of car : 92.700 % Test Accuracy of bird : 82.300 % Test Accuracy of cat : 74.200 % Test Accuracy of deer : 84.800 % Test Accuracy of dog : 79.800 % Test Accuracy of frog : 90.700 % Test Accuracy of horse : 88.600 % Test Accuracy of ship : 92.000 % Test Accuracy of truck : 92.100 % Evaluating Non-Linear Classifier on Convolutional Block 3: Test Accuracy: 82.970 % Accuracy per class: Test Accuracy of plane : 83.700 % Test Accuracy of car : 90.000 % Test Accuracy of bird : 77.300 % Test Accuracy of cat : 73.800 % Test Accuracy of deer : 80.500 % Test Accuracy of dog : 73.900 % Test Accuracy of frog : 87.700 % Test Accuracy of horse : 85.300 % Test Accuracy of ship : 90.300 % Test Accuracy of truck : 87.200 % Evaluating Non-Linear Classifier on Convolutional Block 4: Test Accuracy: 69.830 % Accuracy per class: Test Accuracy of plane : 78.000 % Test Accuracy of car : 73.600 % Test Accuracy of bird : 60.800 % Test Accuracy of cat : 54.600 % Test Accuracy of deer : 67.200 % Test Accuracy of dog : 59.500 % Test Accuracy of frog : 79.400 % Test Accuracy of horse : 73.900 % Test Accuracy of ship : 75.700 % Test Accuracy of truck : 75.600 % Evaluating Non-Linear Classifier on Convolutional Block 5: Test Accuracy: 36.900 % Accuracy per class: Test Accuracy of plane : 44.700 % Test Accuracy of car : 39.400 % Test Accuracy of bird : 29.400 % Test Accuracy of cat : 26.200 % Test Accuracy of deer : 24.300 % Test Accuracy of dog : 31.500 % Test Accuracy of frog : 42.900 % Test Accuracy of horse : 40.200 % Test Accuracy of ship : 47.900 % Test Accuracy of truck : 42.500 % -------------------------------------------------------------------------------- Starting to evaluate Convolutional Classifier: Evaluating Convolutional Classifier on Convolutional Block 1 Test Accuracy: 86.150 % Accuracy per class: Test Accuracy of plane : 89.500 % Test Accuracy of car : 93.000 % Test Accuracy of bird : 79.100 % Test Accuracy of cat : 73.000 % Test Accuracy of deer : 86.400 % Test Accuracy of dog : 79.900 % Test Accuracy of frog : 90.700 % Test Accuracy of horse : 88.300 % Test Accuracy of ship : 90.700 % Test Accuracy of truck : 90.900 % Evaluating Convolutional Classifier on Convolutional Block 2 Test Accuracy: 88.370 % Accuracy per class: Test Accuracy of plane : 88.900 % Test Accuracy of car : 92.700 % Test Accuracy of bird : 84.400 % Test Accuracy of cat : 79.700 % Test Accuracy of deer : 87.900 % Test Accuracy of dog : 81.500 % Test Accuracy of frog : 91.500 % Test Accuracy of horse : 88.900 % Test Accuracy of ship : 94.300 % Test Accuracy of truck : 93.900 % Evaluating Convolutional Classifier on Convolutional Block 3 Test Accuracy: 85.150 % Accuracy per class: Test Accuracy of plane : 86.800 % Test Accuracy of car : 90.700 % Test Accuracy of bird : 79.000 % Test Accuracy of cat : 76.600 % Test Accuracy of deer : 84.900 % Test Accuracy of dog : 76.100 % Test Accuracy of frog : 89.800 % Test Accuracy of horse : 87.400 % Test Accuracy of ship : 90.500 % Test Accuracy of truck : 89.700 % Evaluating Convolutional Classifier on Convolutional Block 4 Test Accuracy: 72.950 % Accuracy per class: Test Accuracy of plane : 78.400 % Test Accuracy of car : 77.300 % Test Accuracy of bird : 63.700 % Test Accuracy of cat : 61.500 % Test Accuracy of deer : 72.900 % Test Accuracy of dog : 65.000 % Test Accuracy of frog : 81.600 % Test Accuracy of horse : 75.700 % Test Accuracy of ship : 76.100 % Test Accuracy of truck : 77.300 % Evaluating Convolutional Classifier on Convolutional Block 5 Test Accuracy: 41.480 % Accuracy per class: Test Accuracy of plane : 49.200 % Test Accuracy of car : 44.200 % Test Accuracy of bird : 33.700 % Test Accuracy of cat : 33.900 % Test Accuracy of deer : 27.800 % Test Accuracy of dog : 35.900 % Test Accuracy of frog : 52.300 % Test Accuracy of horse : 41.400 % Test Accuracy of ship : 50.000 % Test Accuracy of truck : 46.400 %
Out[8]:
{'Accuracy ConvClassifier ConvBlock 1': 86.15,
'Accuracy ConvClassifier ConvBlock 2': 88.37,
'Accuracy ConvClassifier ConvBlock 3': 85.15,
'Accuracy ConvClassifier ConvBlock 4': 72.95,
'Accuracy ConvClassifier ConvBlock 5': 41.48,
'Accuracy Non-Linear ConvBlock 1': 82.99,
'Accuracy Non-Linear ConvBlock 2': 86.61,
'Accuracy Non-Linear ConvBlock 3': 82.97,
'Accuracy Non-Linear ConvBlock 4': 69.83,
'Accuracy Non-Linear ConvBlock 5': 36.9,
'Accuracy Rotation Task': 92.0875,
'Class Accuracy ConvClassifier ConvBlock 1': [89.5,
93.0,
79.1,
73.0,
86.4,
79.9,
90.7,
88.3,
90.7,
90.9],
'Class Accuracy ConvClassifier ConvBlock 2': [88.9,
92.7,
84.4,
79.7,
87.9,
81.5,
91.5,
88.9,
94.3,
93.9],
'Class Accuracy ConvClassifier ConvBlock 3': [86.8,
90.7,
79.0,
76.6,
84.9,
76.1,
89.8,
87.4,
90.5,
89.7],
'Class Accuracy ConvClassifier ConvBlock 4': [78.4,
77.3,
63.7,
61.5,
72.9,
65.0,
81.6,
75.7,
76.1,
77.3],
'Class Accuracy ConvClassifier ConvBlock 5': [49.2,
44.2,
33.7,
33.9,
27.8,
35.9,
52.3,
41.4,
50.0,
46.4],
'Class Accuracy Non-Linear ConvBlock 1': [82.7,
90.2,
77.8,
68.3,
79.1,
73.5,
88.6,
88.0,
91.8,
89.9],
'Class Accuracy Non-Linear ConvBlock 2': [88.9,
92.7,
82.3,
74.2,
84.8,
79.8,
90.7,
88.6,
92.0,
92.1],
'Class Accuracy Non-Linear ConvBlock 3': [83.7,
90.0,
77.3,
73.8,
80.5,
73.9,
87.7,
85.3,
90.3,
87.2],
'Class Accuracy Non-Linear ConvBlock 4': [78.0,
73.6,
60.8,
54.6,
67.2,
59.5,
79.4,
73.9,
75.7,
75.6],
'Class Accuracy Non-Linear ConvBlock 5': [44.7,
39.4,
29.4,
26.2,
24.3,
31.5,
42.9,
40.2,
47.9,
42.5],
'Class Accuracy Rotation Task': [92.38, 92.15, 91.85, 91.97]}
In [13]:
# 5 ConvBlock RotNet model and Classifiers new
ev.evaluate_all(5, testloader, classes)
Evaluating RotNet model with 5 Convolutional Blocks: Evaluating Rotation Task: Test Accuracy: 92.225 % Accuracy per class: Test Accuracy of original : 92.460 % Test Accuracy of 90 rotation : 92.530 % Test Accuracy of 180 rotation : 92.180 % Test Accuracy of 270 rotation : 91.730 % -------------------------------------------------------------------------------- Starting to evaluate Non-Linear Classifier: Evaluating Non-Linear Classifier on Convolutional Block 1: Test Accuracy: 82.640 % Accuracy per class: Test Accuracy of plane : 84.200 % Test Accuracy of car : 90.300 % Test Accuracy of bird : 76.000 % Test Accuracy of cat : 67.700 % Test Accuracy of deer : 81.900 % Test Accuracy of dog : 73.100 % Test Accuracy of frog : 88.000 % Test Accuracy of horse : 86.400 % Test Accuracy of ship : 90.900 % Test Accuracy of truck : 87.900 % Evaluating Non-Linear Classifier on Convolutional Block 2: Test Accuracy: 86.980 % Accuracy per class: Test Accuracy of plane : 88.900 % Test Accuracy of car : 92.400 % Test Accuracy of bird : 83.400 % Test Accuracy of cat : 75.600 % Test Accuracy of deer : 86.200 % Test Accuracy of dog : 80.600 % Test Accuracy of frog : 91.200 % Test Accuracy of horse : 89.300 % Test Accuracy of ship : 92.000 % Test Accuracy of truck : 90.200 % Evaluating Non-Linear Classifier on Convolutional Block 3: Test Accuracy: 83.740 % Accuracy per class: Test Accuracy of plane : 85.100 % Test Accuracy of car : 90.500 % Test Accuracy of bird : 79.400 % Test Accuracy of cat : 71.700 % Test Accuracy of deer : 81.900 % Test Accuracy of dog : 75.500 % Test Accuracy of frog : 88.700 % Test Accuracy of horse : 86.200 % Test Accuracy of ship : 90.100 % Test Accuracy of truck : 88.300 % Evaluating Non-Linear Classifier on Convolutional Block 4: Test Accuracy: 75.130 % Accuracy per class: Test Accuracy of plane : 78.000 % Test Accuracy of car : 79.000 % Test Accuracy of bird : 67.600 % Test Accuracy of cat : 63.000 % Test Accuracy of deer : 73.900 % Test Accuracy of dog : 62.900 % Test Accuracy of frog : 85.800 % Test Accuracy of horse : 80.300 % Test Accuracy of ship : 80.900 % Test Accuracy of truck : 79.900 % Evaluating Non-Linear Classifier on Convolutional Block 5: Test Accuracy: 39.120 % Accuracy per class: Test Accuracy of plane : 45.100 % Test Accuracy of car : 42.400 % Test Accuracy of bird : 29.900 % Test Accuracy of cat : 27.400 % Test Accuracy of deer : 30.900 % Test Accuracy of dog : 33.600 % Test Accuracy of frog : 48.600 % Test Accuracy of horse : 40.700 % Test Accuracy of ship : 48.600 % Test Accuracy of truck : 44.000 % -------------------------------------------------------------------------------- Starting to evaluate Convolutional Classifier: Evaluating Convolutional Classifier on Convolutional Block 1 Test Accuracy: 85.280 % Accuracy per class: Test Accuracy of plane : 87.800 % Test Accuracy of car : 92.700 % Test Accuracy of bird : 79.600 % Test Accuracy of cat : 73.300 % Test Accuracy of deer : 81.900 % Test Accuracy of dog : 77.100 % Test Accuracy of frog : 91.200 % Test Accuracy of horse : 88.100 % Test Accuracy of ship : 91.400 % Test Accuracy of truck : 89.700 % Evaluating Convolutional Classifier on Convolutional Block 2 Test Accuracy: 89.000 % Accuracy per class: Test Accuracy of plane : 90.900 % Test Accuracy of car : 92.700 % Test Accuracy of bird : 85.000 % Test Accuracy of cat : 80.900 % Test Accuracy of deer : 89.600 % Test Accuracy of dog : 82.000 % Test Accuracy of frog : 93.100 % Test Accuracy of horse : 92.400 % Test Accuracy of ship : 91.600 % Test Accuracy of truck : 91.800 % Evaluating Convolutional Classifier on Convolutional Block 3 Test Accuracy: 85.400 % Accuracy per class: Test Accuracy of plane : 85.800 % Test Accuracy of car : 89.800 % Test Accuracy of bird : 80.300 % Test Accuracy of cat : 76.000 % Test Accuracy of deer : 87.700 % Test Accuracy of dog : 76.800 % Test Accuracy of frog : 90.500 % Test Accuracy of horse : 88.000 % Test Accuracy of ship : 90.700 % Test Accuracy of truck : 88.400 % Evaluating Convolutional Classifier on Convolutional Block 4 Test Accuracy: 76.960 % Accuracy per class: Test Accuracy of plane : 78.200 % Test Accuracy of car : 78.500 % Test Accuracy of bird : 72.200 % Test Accuracy of cat : 63.700 % Test Accuracy of deer : 78.400 % Test Accuracy of dog : 68.800 % Test Accuracy of frog : 85.700 % Test Accuracy of horse : 78.700 % Test Accuracy of ship : 83.500 % Test Accuracy of truck : 81.900 % Evaluating Convolutional Classifier on Convolutional Block 5 Test Accuracy: 43.900 % Accuracy per class: Test Accuracy of plane : 56.000 % Test Accuracy of car : 44.100 % Test Accuracy of bird : 38.400 % Test Accuracy of cat : 32.800 % Test Accuracy of deer : 33.400 % Test Accuracy of dog : 36.300 % Test Accuracy of frog : 51.400 % Test Accuracy of horse : 46.400 % Test Accuracy of ship : 48.100 % Test Accuracy of truck : 52.100 %
Out[13]:
{'Accuracy ConvClassifier ConvBlock 1': 85.28,
'Accuracy ConvClassifier ConvBlock 2': 89.0,
'Accuracy ConvClassifier ConvBlock 3': 85.4,
'Accuracy ConvClassifier ConvBlock 4': 76.96,
'Accuracy ConvClassifier ConvBlock 5': 43.9,
'Accuracy Non-Linear ConvBlock 1': 82.64,
'Accuracy Non-Linear ConvBlock 2': 86.98,
'Accuracy Non-Linear ConvBlock 3': 83.74,
'Accuracy Non-Linear ConvBlock 4': 75.13,
'Accuracy Non-Linear ConvBlock 5': 39.12,
'Accuracy Rotation Task': 92.225,
'Class Accuracy ConvClassifier ConvBlock 1': [87.8,
92.7,
79.6,
73.3,
81.9,
77.1,
91.2,
88.1,
91.4,
89.7],
'Class Accuracy ConvClassifier ConvBlock 2': [90.9,
92.7,
85.0,
80.9,
89.6,
82.0,
93.1,
92.4,
91.6,
91.8],
'Class Accuracy ConvClassifier ConvBlock 3': [85.8,
89.8,
80.3,
76.0,
87.7,
76.8,
90.5,
88.0,
90.7,
88.4],
'Class Accuracy ConvClassifier ConvBlock 4': [78.2,
78.5,
72.2,
63.7,
78.4,
68.8,
85.7,
78.7,
83.5,
81.9],
'Class Accuracy ConvClassifier ConvBlock 5': [56.0,
44.1,
38.4,
32.8,
33.4,
36.3,
51.4,
46.4,
48.1,
52.1],
'Class Accuracy Non-Linear ConvBlock 1': [84.2,
90.3,
76.0,
67.7,
81.9,
73.1,
88.0,
86.4,
90.9,
87.9],
'Class Accuracy Non-Linear ConvBlock 2': [88.9,
92.4,
83.4,
75.6,
86.2,
80.6,
91.2,
89.3,
92.0,
90.2],
'Class Accuracy Non-Linear ConvBlock 3': [85.1,
90.5,
79.4,
71.7,
81.9,
75.5,
88.7,
86.2,
90.1,
88.3],
'Class Accuracy Non-Linear ConvBlock 4': [78.0,
79.0,
67.6,
63.0,
73.9,
62.9,
85.8,
80.3,
80.9,
79.9],
'Class Accuracy Non-Linear ConvBlock 5': [45.1,
42.4,
29.9,
27.4,
30.9,
33.6,
48.6,
40.7,
48.6,
44.0],
'Class Accuracy Rotation Task': [92.46, 92.53, 92.18, 91.73]}
In [6]:
# Supervised NIN
ev.evaluate_all(0, testloader, classes)
Evaluating Supervised NIN Classification Task: Test Accuracy: 91.390 % Test Accuracy of plane : 91.600 % Test Accuracy of car : 95.900 % Test Accuracy of bird : 87.000 % Test Accuracy of cat : 83.200 % Test Accuracy of deer : 92.100 % Test Accuracy of dog : 85.500 % Test Accuracy of frog : 94.800 % Test Accuracy of horse : 93.900 % Test Accuracy of ship : 94.700 % Test Accuracy of truck : 95.200 %
Out[6]:
{'Accuracy Supervised NIN': 91.39,
'Class Accuracy Supervised NIN': [91.6,
95.9,
87.0,
83.2,
92.1,
85.5,
94.8,
93.9,
94.7,
95.2]}
In [6]:
# semi-supervised
ev.evaluate_all(-1, testloader, classes, semi=[20, 100, 400, 1000, 5000])
Evaluating Semi-supervised Experiment with 20 images per class: Test Accuracy: 62.760 % Test Accuracy of plane : 63.600 % Test Accuracy of car : 74.300 % Test Accuracy of bird : 46.500 % Test Accuracy of cat : 37.200 % Test Accuracy of deer : 57.000 % Test Accuracy of dog : 59.400 % Test Accuracy of frog : 76.900 % Test Accuracy of horse : 66.000 % Test Accuracy of ship : 71.700 % Test Accuracy of truck : 75.000 % Evaluating supervised NIN Experiment with 20 images per class: Test Accuracy: 31.570 % Test Accuracy of plane : 38.400 % Test Accuracy of car : 42.600 % Test Accuracy of bird : 17.300 % Test Accuracy of cat : 24.300 % Test Accuracy of deer : 23.900 % Test Accuracy of dog : 28.200 % Test Accuracy of frog : 35.300 % Test Accuracy of horse : 36.000 % Test Accuracy of ship : 37.400 % Test Accuracy of truck : 32.300 % -------------------------------------------------------------------------------- Evaluating Semi-supervised Experiment with 100 images per class: Test Accuracy: 71.780 % Test Accuracy of plane : 70.200 % Test Accuracy of car : 82.800 % Test Accuracy of bird : 58.300 % Test Accuracy of cat : 52.200 % Test Accuracy of deer : 67.100 % Test Accuracy of dog : 67.000 % Test Accuracy of frog : 81.000 % Test Accuracy of horse : 75.900 % Test Accuracy of ship : 80.300 % Test Accuracy of truck : 83.000 % Evaluating supervised NIN Experiment with 100 images per class: Test Accuracy: 47.100 % Test Accuracy of plane : 49.400 % Test Accuracy of car : 61.000 % Test Accuracy of bird : 29.700 % Test Accuracy of cat : 30.000 % Test Accuracy of deer : 37.200 % Test Accuracy of dog : 39.800 % Test Accuracy of frog : 50.000 % Test Accuracy of horse : 54.100 % Test Accuracy of ship : 64.900 % Test Accuracy of truck : 54.900 % -------------------------------------------------------------------------------- Evaluating Semi-supervised Experiment with 400 images per class: Test Accuracy: 80.190 % Test Accuracy of plane : 81.800 % Test Accuracy of car : 89.500 % Test Accuracy of bird : 72.900 % Test Accuracy of cat : 61.600 % Test Accuracy of deer : 79.700 % Test Accuracy of dog : 72.500 % Test Accuracy of frog : 86.800 % Test Accuracy of horse : 83.800 % Test Accuracy of ship : 85.700 % Test Accuracy of truck : 87.600 % Evaluating supervised NIN Experiment with 400 images per class: Test Accuracy: 71.120 % Test Accuracy of plane : 71.500 % Test Accuracy of car : 81.700 % Test Accuracy of bird : 56.800 % Test Accuracy of cat : 53.200 % Test Accuracy of deer : 67.200 % Test Accuracy of dog : 62.000 % Test Accuracy of frog : 77.800 % Test Accuracy of horse : 76.700 % Test Accuracy of ship : 82.900 % Test Accuracy of truck : 81.400 % -------------------------------------------------------------------------------- Evaluating Semi-supervised Experiment with 1000 images per class: Test Accuracy: 83.670 % Test Accuracy of plane : 83.900 % Test Accuracy of car : 91.600 % Test Accuracy of bird : 77.900 % Test Accuracy of cat : 69.300 % Test Accuracy of deer : 83.200 % Test Accuracy of dog : 78.600 % Test Accuracy of frog : 88.100 % Test Accuracy of horse : 87.700 % Test Accuracy of ship : 88.500 % Test Accuracy of truck : 87.900 % Evaluating supervised NIN Experiment with 1000 images per class: Test Accuracy: 81.370 % Test Accuracy of plane : 82.500 % Test Accuracy of car : 90.200 % Test Accuracy of bird : 73.900 % Test Accuracy of cat : 64.200 % Test Accuracy of deer : 78.000 % Test Accuracy of dog : 75.000 % Test Accuracy of frog : 86.600 % Test Accuracy of horse : 85.100 % Test Accuracy of ship : 90.600 % Test Accuracy of truck : 87.600 % -------------------------------------------------------------------------------- Evaluating Semi-supervised Experiment with 5000 images per class: Test Accuracy: 88.940 % Test Accuracy of plane : 89.000 % Test Accuracy of car : 92.200 % Test Accuracy of bird : 86.100 % Test Accuracy of cat : 79.200 % Test Accuracy of deer : 88.700 % Test Accuracy of dog : 83.300 % Test Accuracy of frog : 92.500 % Test Accuracy of horse : 91.200 % Test Accuracy of ship : 93.700 % Test Accuracy of truck : 93.500 % Evaluating supervised NIN Experiment with 5000 images per class: Test Accuracy: 91.370 % Test Accuracy of plane : 92.100 % Test Accuracy of car : 95.900 % Test Accuracy of bird : 87.800 % Test Accuracy of cat : 81.200 % Test Accuracy of deer : 92.100 % Test Accuracy of dog : 86.900 % Test Accuracy of frog : 94.100 % Test Accuracy of horse : 93.300 % Test Accuracy of ship : 95.400 % Test Accuracy of truck : 94.900 % --------------------------------------------------------------------------------
Out[6]:
{'Accuracy Semi-supervised 100': 71.78,
'Accuracy Semi-supervised 1000': 83.67,
'Accuracy Semi-supervised 20': 62.76,
'Accuracy Semi-supervised 400': 80.19,
'Accuracy Semi-supervised 5000': 88.94,
'Accuracy Supervised NIN 100': 47.1,
'Accuracy Supervised NIN 1000': 81.37,
'Accuracy Supervised NIN 20': 31.57,
'Accuracy Supervised NIN 400': 71.12,
'Accuracy Supervised NIN 5000': 91.37,
'Class Accuracy Semi-supervised 100': [70.2,
82.8,
58.3,
52.2,
67.1,
67.0,
81.0,
75.9,
80.3,
83.0],
'Class Accuracy Semi-supervised 1000': [83.9,
91.6,
77.9,
69.3,
83.2,
78.6,
88.1,
87.7,
88.5,
87.9],
'Class Accuracy Semi-supervised 20': [63.6,
74.3,
46.5,
37.2,
57.0,
59.4,
76.9,
66.0,
71.7,
75.0],
'Class Accuracy Semi-supervised 400': [81.8,
89.5,
72.9,
61.6,
79.7,
72.5,
86.8,
83.8,
85.7,
87.6],
'Class Accuracy Semi-supervised 5000': [89.0,
92.2,
86.1,
79.2,
88.7,
83.3,
92.5,
91.2,
93.7,
93.5],
'Class Accuracy Supervised NIN 100': [49.4,
61.0,
29.7,
30.0,
37.2,
39.8,
50.0,
54.1,
64.9,
54.9],
'Class Accuracy Supervised NIN 1000': [82.5,
90.2,
73.9,
64.2,
78.0,
75.0,
86.6,
85.1,
90.6,
87.6],
'Class Accuracy Supervised NIN 20': [38.4,
42.6,
17.3,
24.3,
23.9,
28.2,
35.3,
36.0,
37.4,
32.3],
'Class Accuracy Supervised NIN 400': [71.5,
81.7,
56.8,
53.2,
67.2,
62.0,
77.8,
76.7,
82.9,
81.4],
'Class Accuracy Supervised NIN 5000': [92.1,
95.9,
87.8,
81.2,
92.1,
86.9,
94.1,
93.3,
95.4,
94.9]}
Plots¶
In [7]:
p.plot_all([20, 100, 400, 1000, 5000])
